s3-ap-southeast-1.amazonaws.com... · Declaration I declare that the thesis entitled "Solution of...
Transcript of s3-ap-southeast-1.amazonaws.com... · Declaration I declare that the thesis entitled "Solution of...
SOLUTION OF HIGHER ORDERDIFFERENTIAL EQUATIONS WITHFUZZY INITIAL AND BOUNDARY
CONDITIONS
A Thesis submitted to Gujarat Technological University
for the award of
Doctor of Philosophyin
Science - Mathematicsby
Patel Komalben RameshbhaiEnrollment No.: 149997673006
under supervision of
Dr. Narendrasinh B. Desai
GUJARAT TECHNOLOGICAL UNIVERSITY
AHMEDABAD
SEPTEMBER 2020
SOLUTION OF HIGHER ORDERDIFFERENTIAL EQUATIONS WITHFUZZY INITIAL AND BOUNDARY
CONDITIONS
A Thesis submitted to Gujarat Technological University
for the award of
Doctor of Philosophyin
Science - Mathematicsby
Patel Komalben RameshbhaiEnrollment No.: 149997673006
under supervision of
Dr. Narendrasinh B. Desai
GUJARAT TECHNOLOGICAL UNIVERSITY
AHMEDABAD
SEPTEMBER 2020
c©Patel Komalben Rameshbhai
ii
DeclarationI declare that the thesis entitled "Solution of Higher Order Differential Equations
with Fuzzy Initial and Boundary Conditions" submitted by me for the degree of
Doctor of Philosophy is the record of research work carried out by me during the period
from March 2015 to September 2019 under the supervision of Dr. Narendrasinh B.
Desai, Head of Mathematics Department, A. D. Patel Institute of Technology, New V. V.
Nagar, Anand, and this has not formed the basis for the award of any degree, diploma,
associateship, fellowship, titles in this or any other university or other institution of
higher learning. I further declare that the material obtained from other sources has been
duly acknowledged in the thesis. I shall be solely responsible for any plagiarism or
other irregularities, if noticed in the thesis.
Signature of the Research Scholar:
Date: 23/09/2020
Name of Research Scholar: Patel Komalben RameshbhaiPlace: Vadodara.
iii
CertificateI certify that the work incorporated in the thesis "Solution of Higher Order
Differential Equations with Fuzzy Initial and Boundary Conditions" submitted by
Patel Komalben Rameshbhai was carried out by the candidate under my guidance.
To the best of my knowledge: (i) the candidate has not submitted the same research
work to any other institution for any degree, diploma, associateship, fellowship or
other similar titles; (ii) the thesis submitted is a record of original research work done
by Research Scholar during the period of study under my supervision; and (iii) the
thesis represents independent research work on the part of the Research Scholar.
Signature of Supervisor:
Date: 23/09/2020
Name of Supervisor: Dr. Narendrasinh B. DesaiPlace: New V. V. Nagar.
iv
Course-work Completion CertificateThis is to certify that Patel Komalben Rameshbhai, enrolment no. 149997673006 is
a PhD scholar enrolled for PhD program in the branch Science-Mathematics of
Gujarat Technological University, Ahmedabad.
(Please tick the relevant option(s))
f She has been exempted from the course-work (successfully completed during
M.Phil Course)
f She has been exempted from Research Methodology Course only (successfully
completed during M.Phil Course)
f! She has successfully completed the PhD course work for the partial requirement
for the award of PhD Degree. Her performance in the course work is as follows-
Grade Obtained in Research Methodology Grade Obtained in Self Study Course (Core Subject)
(PH001) (PH002)
BC AB
Supervisor’s Sign:
Name of Supervisor: Dr. Narendrasinh B. Desai
v
Originality Report CertificateIt is certified that PhD Thesis entitled "Solution of Higher Order Differential
Equations with Fuzzy Initial and Boundary Conditions" by Patel Komalben
Rameshbhai has been examined by us. We undertake the following:
a. Thesis has significant new work / knowledge as compared to already published or
are under consideration to be published elsewhere. No sentence, equation, diagram,
table, paragraph or section has been copied verbatim from previous work unless it is
placed under quotation marks and duly referenced.
b. The work presented is original and own work of the author (i.e. there is no
plagiarism). No ideas, processes, results or words of others have been presented as
author’s own work.
c. There is no fabrication of data or results which have been compiled / analyzed.
d. There is no falsification by manipulating research materials, equipment or
processes, or changing or omitting data or results such that the research is not
accurately represented in the research record.
e. The thesis has been checked using Turnitin (copy of originality report attached)
and found within limits as per GTU Plagiarism Policy and instructions issued from
time to time (i.e. permitted similarity index ≤ 10%).
Signature of the Research Scholar: Date: 23/09/2020
Name of Research Scholar: Patel Komalben RameshbhaiPlace: Vadodara.
Signature of Supervisor: Date: 23/09/2020
Name of Supervisor: Dr. Narendrasinh B. DesaiPlace: New V. V. Nagar.
vi
vii
PhD THESIS Non-Exclusive License to
GUJARAT TECHNOLOGICAL
UNIVERSITYIn consideration of being a PhD Research Scholar at GTU and in the interests of the
facilitation of research at GTU and elsewhere, I, Patel Komalben Rameshbhai, having
enrolment no. 149997673006, hereby grant a non-exclusive, royalty free and perpetual
license to GTU on the following terms:
a). GTU is permitted to archive, reproduce and distribute my thesis, in whole or in
part, and/or my abstract, in whole or in part (referred to collectively as the "Work")
anywhere in the world, for non-commercial purposes, in all forms of media;
b). GTU is permitted to authorize, sub-lease, sub-contract or procure any of the acts
mentioned in paragraph (a);
c). GTU is authorized to submit the Work at any National / International Library, under
the authority of their "Thesis Non-Exclusive License";
d). The Universal Copyright Notice ( c©) shall appear on all copies made under the
authority of this license;
e). I undertake to submit my thesis, through my University, to any Library and
Archives. Any abstract submitted with the thesis will be considered to form part of
the thesis.
f). I represent that my thesis is my original work, does not infringe any rights of others,
including privacy rights, and that I have the right to make the grant conferred by this
non-exclusive license.
g). If third party copyrighted material was included in my thesis for which, under the
terms of the Copyright Act, written permission from the copyright owners is
required, I have obtained such permission from the copyright owners to do the acts
mentioned in paragraph (a) above for the full term of copyright protection.
viii
h). I retain copyright ownership and moral rights in my thesis, and may deal with the
copyright in my thesis, in any way consistent with rights granted by me to my
University in this non-exclusive license.
i). I further promise to inform any person to whom I may hereafter assign or license
my copyright in my thesis of the rights granted by me to my University in this
non-exclusive license.
j). I am aware of and agree to accept the conditions and regulations of PhD including
all policy matters related to authorship and plagiarism.
Signature of the Research Scholar:
Name of Research Scholar:Patel Komalben RameshbhaiDate: 23/09/2020 Place: Vadodara.
Signature of Supervisor:
Name of Supervisor: Dr. Narendrasinh B. DesaiDate: 23/09/2020 Place: New V. V. Nagar.
Seal:
ix
Thesis Approval FormThe viva-voce of the PhD Thesis submitted by Patel Komalben Rameshbhai
(enrolment no.149997673006) entitled "Solution of Higher Order Differential
Equations with Fuzzy Initial and Boundary Conditions" was conducted on
Wednesday, 23 September 2020 at Gujarat Technological University.
(Please tick any one of the following option)
f! The performance of the candidate was satisfactory. We recommend that she be
awarded the PhD degree.
f Any further modifications in research work recommended by the panel after 3
months from the date of first viva-voce upon request of the Supervisor or request
of Independent Research Scholar after which viva-voce can be re-conducted by
the same panel again.
(briefly specify the modifications suggested by the panel)
f The performance of the candidate was unsatisfactory. We recommend that he/she
should not be awarded the PhD degree.
(The panel must give justifications for rejecting the research work)
Dr. Narendrasinh B. Desai
Dr. Pragnesh Gajjar
Name and Signature of Supervisor with Seal 1)(External Examiner:1) Name and Signature
Dr. Shrikant Chaudhari2)(External Examiner:2) Name and Signature 3)(External Examiner:3) Name and Signature
x
AbstractThe study of the solution of higher-order differential equations(DEs) with fuzzy initial
and boundary conditions(FIBCs) and the development of the solution methods for
higher-order differential equations are the best practical solutions for real-life
problems. The solution of higher-order DEs with FIBCs is one of the most important
concepts in many areas of applied sciences and engineering such as, mechanical
engineering, civil engineering, electrical engineering, electrical and communication
engineering, environmental engineering, chemical engineering, medical science, etc.
In recent years, extensive research have been carried out to study the solution of
higher-order DEs with FIBCs. This field has achieved extensive attention due to its
broad range of applications in the field of science, engineering and real-life practical
problems. In particular, finding the solution of higher-order differential equation with
fuzzy initial and boundary conditions is a central problem in the various fields of
engineering and sciences. The scope of the present study lies in increasing importance
of the solution of higher-order fuzzy differential equations(FDEs) as well as fuzziness
in the forcing function. Thus, an investigation of such research area leads directly or
indirectly to studying another research area far away from the first in terms of physical
context, yet related through the fundamental principles of the ordinary differential
equation(ODE) that involved fuzziness in initial and boundary conditions. Due to the
vast scope of application of DEs in the fields of engineering and sciences, the specific
problems are almost unlimited in fields that involved fuzziness and therefore it is
reasonable to select such types of problems for discussion here. Accordingly, a
selection of more interesting problems of current interest has been made for
mathematical treatment in the thesis. The problems investigated in the present study
are concerned with the application of a solution of higher-order DEs that contain
fuzziness in initial and boundary conditions as well as in forcing function that are
occurring in different fields of engineering and sciences.
This work has been devoted to the study of higher-order DEs with FIBCs by using
suitable methods. Here, we have discussed fuzzy dynamic problems like the bending of
an elastic beam under the constant load in civil engineering, Ariy’s non-homogeneous
xi
differential equation in optics and the simple pendulum problem in physics, the mass-
spring system in mechanical engineering, etc. The mathematical problem of different
physical phenomenon gives us one-dimensional differential equations with fuzzy initial
and boundary conditions. This Ph.D. work would be useful for solving fuzzy dynamic
problems arising in different fields of engineering and sciences.
xii
Acknowledgement
It is a matter of tremendous pleasure for me to submit my Ph.D. thesis entitled "Solutionof Higher Order Differential Equations with Fuzzy Initial and Boundary Conditions" inthe subject of Mathematics. My registration was done in the year 2015 for carrying outthe work related to the subject which seemed to me to be herculean task ab initio butwith the passing of time, everything seemed to be within the reach by god’s grace.
I feel highly indebted to my guide Dr. Narendrasinh B. Desai, Head, Departmentof Mathematics, A. D. Patel Institute of Technology, New V. V. Nagar, Anand. He hasbeen a consistent supporter during the course of my work. He has been the real guidingforce behind this work who inspired me to go on with undaunted courage. My DoctorateProgress Committee members also guided me through all these years. I would like tothank Dr. J. M. Dhodiya, Associate Professor, SVNIT, Surat, and Dr. M. S. Joshi,Associate Professor, C. K. Pithawala College of Engineering & Technology, Surat, whohave reviewed my research work time to time and given the valuable suggestion in myresearch work.
During the process of preparation of my thesis, I have referred many books andresearch papers for which I personally thank the authors who provided me a deep insightand inspiration for the solution of problems.
I sincerely thank my Director Dr. K baba Pai and my Dean Academics Dr. S. K. Vijand Dr.Som Sahani Head, Mathematics Department, who have allowed and cooperatedto do this research work as a part-time candidate along with my job responsibilities. Iwould like to thanks to my Ph.D. colleagues like Dr. D. J. Prajapati, Dr. M. A. Pateland Mr. Ujjaval Trivedi, AHOD of Mechanical department as well as all the facultymembers of the Mathematics department of ITM(SLS) Baroda University, Vadodara.
No value is more precious than the value of the family hence, I sincerely thank myin-laws, my husband Hiren Patel and my son Kavya Patel for the great help and constantinspiration throughout my work. They have done everything they could to make me feelrelaxed and inspired.
At the end, I would like to appreciate all the help and support extended to meduring this journey.
Komal R. Patel
xiii
Contents
Abstract xi
Acknowledgement xiii
List of Abbreviations xvii
List of Symbols xviii
List of Figures xx
List of Tables xxii
1 General Introduction 11.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.2 Structure of the thesis . . . . . . . . . . . . . . . . . . . . . . . . . . . 21.3 Brief description on the state of the art of the research topic . . . . . . . 31.4 Definition of the problem . . . . . . . . . . . . . . . . . . . . . . . . . 51.5 Objective and future scope of work . . . . . . . . . . . . . . . . . . . . 51.6 Original contribution by the thesis . . . . . . . . . . . . . . . . . . . . 61.7 Methodology of Research and Results/Comparisons . . . . . . . . . . . 61.8 Achievements with respect to objectives . . . . . . . . . . . . . . . . . 71.9 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2 Fundamentals of Fuzzy Set Theory and Basic Concepts 92.1 History of development of fuzzy set theory . . . . . . . . . . . . . . . . 92.2 Crisp set vs. Fuzzy set . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.2.1 Crisp set . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102.2.2 Fuzzy set . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.3 Standard operations on fuzzy sets: . . . . . . . . . . . . . . . . . . . . 112.3.1 Logical connectives . . . . . . . . . . . . . . . . . . . . . . . . 11
2.4 Fuzzy set and Membership function . . . . . . . . . . . . . . . . . . . 132.4.1 Membership function . . . . . . . . . . . . . . . . . . . . . . . 13
2.5 Properties of Membership function . . . . . . . . . . . . . . . . . . . . 132.5.1 Core . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142.5.2 Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142.5.3 Boundary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142.5.4 Height . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152.5.5 Normal fuzzy set . . . . . . . . . . . . . . . . . . . . . . . . . 152.5.6 Subnormal fuzzy set . . . . . . . . . . . . . . . . . . . . . . . 162.5.7 Convex fuzzy set . . . . . . . . . . . . . . . . . . . . . . . . . 16
xiv
Contents xv
2.5.8 r-cut . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172.6 Fuzzy number . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.6.1 Triangular fuzzy number . . . . . . . . . . . . . . . . . . . . . 182.6.2 Trapezoidal fuzzy number . . . . . . . . . . . . . . . . . . . . 192.6.3 Gaussian fuzzy number . . . . . . . . . . . . . . . . . . . . . . 20
2.7 Fuzzy arithmetic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212.8 Zadeh’s extension principle . . . . . . . . . . . . . . . . . . . . . . . . 212.9 Continuity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222.10 Hukuhara difference . . . . . . . . . . . . . . . . . . . . . . . . . . . 222.11 Hukuhara differential . . . . . . . . . . . . . . . . . . . . . . . . . . . 232.12 Generalized Hukuhara differentiability . . . . . . . . . . . . . . . . . . 232.13 Fuzzy Reimann integration . . . . . . . . . . . . . . . . . . . . . . . . 23
3 Solutions of nth order Differential Equations with Fuzzy Initial Conditionsby using the Properties of Linear Transformations 253.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253.2 Linear transformation . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
3.2.1 Wronskain . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263.3 Problem formulation . . . . . . . . . . . . . . . . . . . . . . . . . . . 263.4 Method of solving nth order differential equation with fuzzy initial
conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273.5 Solution of a higher order differential equation with fuzzy initial condition 30
3.5.1 Solution of a higher order constant coefficient differentialequation with fuzzy initial condition . . . . . . . . . . . . . . . 30
3.5.2 Solution of higher order Cauchy-Euler’s variable coefficientdifferential equations with fuzzy initial conditions . . . . . . . 37
3.6 Result and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . 433.7 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
4 Solutions of the nth order Differential Equations with Fuzzy Initial andBoundary Conditions by using the Gauss Elimination Method 454.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 454.2 Problem formulation . . . . . . . . . . . . . . . . . . . . . . . . . . . 464.3 Method for solving the nth order differential equation with FBCs . . . . 464.4 Solutions of higher order differential equations with fuzzy initial and
boundary conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 554.4.1 Solutions of a second order differential equation with fuzzy
boundary conditions . . . . . . . . . . . . . . . . . . . . . . . 554.4.2 Solution of fourth order differential equation with fuzzy initial
and boundary conditions . . . . . . . . . . . . . . . . . . . . . 594.4.2.1 Example based on the application of differential
equation in civil engineering . . . . . . . . . . . . . 594.5 Result and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . 644.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
Contents xvi
5 Solutions of Higher Order Differential Equations with Fuzzy BoundaryConditions by Finite Difference Method 665.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 665.2 Problem formulation . . . . . . . . . . . . . . . . . . . . . . . . . . . 675.3 Method for solving higher order differential equation with FBCs . . . . 705.4 Solution of the higher order differential equation with fuzzy boundary
conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 725.4.1 Solution of the second order differential equation with fuzzy
boundary conditions . . . . . . . . . . . . . . . . . . . . . . . 725.4.2 Airy’s functions . . . . . . . . . . . . . . . . . . . . . . . . . 77
5.4.2.1 Solution of Airy’s non-homogeneous differentialequation with fuzzy boundary conditions . . . . . . . 77
5.5 Result and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . 845.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
6 Methods for solving Higher Order Differential Equations with Fuzzy Initialand Boundary Conditions by using Laplace Transforms 856.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 856.2 Laplace transform . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
6.2.1 Conditions for the existence of Laplace transforms . . . . . . . 866.3 Strongly generalized Hukuhara differentiability . . . . . . . . . . . . . 866.4 Problem formulation . . . . . . . . . . . . . . . . . . . . . . . . . . . 876.5 Method for solving higher order differential equations with fuzzy initial
conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 886.6 Solution of higher order differential equations with fuzzy initial and
boundary conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 906.6.1 Solution of second order constant coefficient differential
equations with fuzzy initial conditions . . . . . . . . . . . . . . 906.6.1.1 An example based on the application of differential
equation in mechanical engineering . . . . . . . . . . 906.6.1.2 Example based on the application of differential
equation in physics . . . . . . . . . . . . . . . . . . 946.6.2 Solution of second order variable coefficient differential
equations with fuzzy initial and boundary conditions . . . . . . 986.7 Result and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . 1036.8 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
References 104
List of Publications 109
List of Abbreviations
Acronym What (it) Stands ForDE Differential EquationODE Ordinary Differential EquationFDE Fuzzy Differential EquationMF Membership FunctionTMF Triangular Membership FunctionTraMF Trarapezoidal Membership FunctionGMF Gaussian Membership FunctionTFN Triangular Fuzzy NumberTraFN Trarapezoidal Fuzzy NumberGFN Gaussian Fuzzy NumberHD Hukuhara DifferentialGHD Generalized Hukuhara DifferentiabilitySGHD Strongly Generalized Hukuhara DifferentiabilityIVP Initial Value ProblemBVP Boundary Value ProblemFIVP Fuzzy Initial Value ProblemFBVP Fuzzy Boundary Value ProblemFIC Fuzzy Initial ConditionFBC Fuzzy Boundary ConditionFIBC Fuzzy Initial and Boundary ConditionLT Laplace TransformFLT Fuzzy Laplace TransformFRI Fuzzy Reimann IntegrableGEM Gauss Elimination MethodFDM Finite Difference Method
xvii
List of Symbols
symbol nameA crisp setχ characteristic function for crisp setIR set of real numberA fuzzy setE set of real fuzzy numberµ membership function for fuzzy set∨ or ∪ Disjunction or Union∧ or ∩ Conjunction or IntersectionAr r - cutAr+ strong r - cut Hukuhara differenceα Modal valueσl left fuzzinessσr right fuzziness∃ there exist at least∈ member of6= not equal to∀ all elements of the setW WronskainC(n−1)(−∞,∞) the set of n−1 time continuous functionT Linear Transformationy lower-bound of solutiony upper-bound of solutionn number of equal size sub-intervalh step sizeL Laplace transformL length of bridgeE Younge’s modulus of elasticity of material of bridgeI beam cross sectional moment of inertia about horizontal axisEI measurement of beam stiffnessw0 constant loadm mass of springb damping coefficientk spring constant
xviii
List of Symbols xix
θ displacement functionl length of pendulumm1 mass of bobλ constantY exact solutiony approximate solutiony′
first order derivativey′′
second order derivativey′′′
third order derivativeyiv fourth order derivativeyn nth order derivativeα, β fuzzy boundary valuesγk,σak,σbk,δk,ζk,ξk parameters used for matrices entriesζ+,ζ−,ζ ,η parameters used in Taylor’s series
List of Figures
2.1 Lotfi Zadeh [https://www.slideserve.com/tambre/lotfi-zadeh] . . . . . . 92.2 Crisp set vs. Fuzzy set . . . . . . . . . . . . . . . . . . . . . . . . . . 102.3 A . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112.4 B . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122.5 A∧ B . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122.6 A∨ B . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122.7 A
′. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.8 Core, support, and boundaries of a fuzzy set . . . . . . . . . . . . . . . 152.9 Normal fuzzy set . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152.10 Subnormal fuzzy set . . . . . . . . . . . . . . . . . . . . . . . . . . . 162.11 Convex fuzzy set . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172.12 Triangular fuzzy number . . . . . . . . . . . . . . . . . . . . . . . . . 192.13 Trapezoidal fuzzy number . . . . . . . . . . . . . . . . . . . . . . . . 202.14 Gaussian fuzzy number . . . . . . . . . . . . . . . . . . . . . . . . . . 21
3.1 The fuzzy solution obtained by the properties of linear transformations . 313.2 The fuzzy solution obtained by the properties of linear transformations
and the red line represents the crisp solution . . . . . . . . . . . . . . . 323.3 The fuzzy solution obtained by the properties of linear transformation . 343.4 The fuzzy solution obtained by the properties of linear transformations,
and the red line represents the crisp solution . . . . . . . . . . . . . . . 363.5 If t increases, fuzziness disappears . . . . . . . . . . . . . . . . . . . . 363.6 The fuzzy solution obtained by the properties of linear transformations . 383.7 The fuzzy solution obtained by the properties of linear transformations,
The red line represents the crisp solution . . . . . . . . . . . . . . . . . 383.8 The fuzzy solution obtained by properties of linear transformation . . . 403.9 The fuzzy solution obtained by properties of linear transformation, The
red line represents the crisp solution . . . . . . . . . . . . . . . . . . . 413.10 If we increase t fuzziness increases . . . . . . . . . . . . . . . . . . . . 413.11 Mass tapped periodically with a hammer . . . . . . . . . . . . . . . . . 423.12 The fuzzy solution obtained by the properties of linear transformation,
The red line represents the crisp solution . . . . . . . . . . . . . . . . . 43
4.1 The lower-bound and upper-bound for exact solution . . . . . . . . . . 564.2 The lower-bound and upper-bound for approximate solution . . . . . . 594.3 Elastic or Deflection curve . . . . . . . . . . . . . . . . . . . . . . . . 604.4 The lower-bound and upper-bound for exact solution . . . . . . . . . . 614.5 The lower-bound and upper-bound for approximate solution . . . . . . 64
5.1 The lower-bounds and the upper-bounds for an approximate solution . . 765.2 The lower-bounds and the upper-bounds for an approximate solution . . 83
xx
List of Figures xxi
6.1 The mass-spring system . . . . . . . . . . . . . . . . . . . . . . . . . . 916.2 The lower-bound and upper-bound of solution at r = 0 . . . . . . . . . . 926.3 The lower-bound and upper-bound of solution at r = 0.5 . . . . . . . . . 936.4 The lower-bound and upper-bound of solution at r = 0.8 . . . . . . . . . 936.5 The lower-bound and upper-bound of solution at r = 0.9 . . . . . . . . . 936.6 The lower-bound and upper-bound of solution at r = 1 . . . . . . . . . . 936.7 The simple pendulum . . . . . . . . . . . . . . . . . . . . . . . . . . . 956.8 The lower-bound and upper-bound of solution at r = 0 . . . . . . . . . . 966.9 The lower-bound and upper-bound of solution at r = 0.5 . . . . . . . . . 976.10 The lower-bound and upper-bound of solution at r = 0.8 . . . . . . . . . 976.11 The lower-bound and upper-bound of solution at r = 0.9 . . . . . . . . . 976.12 The lower-bound and upper-bound of solution at r = 1 . . . . . . . . . . 976.13 Solution by considering parameters TFN . . . . . . . . . . . . . . . . . 986.14 Solution by considering parameters TraFN . . . . . . . . . . . . . . . . 996.15 Solution by considering parameters GFN . . . . . . . . . . . . . . . . . 1006.16 Solution by considering parameters TFN . . . . . . . . . . . . . . . . . 1016.17 Solution by considering parameters TraFN . . . . . . . . . . . . . . . . 1026.18 Solution by considering parameters GFN . . . . . . . . . . . . . . . . . 102
List of Tables
4.1 The exact values for lower-bounds. . . . . . . . . . . . . . . . . . . . . 564.2 The exact values for upper-bounds. . . . . . . . . . . . . . . . . . . . . 564.3 The approximate values for lower-bounds. . . . . . . . . . . . . . . . . 584.4 The approximate values for upper-bounds. . . . . . . . . . . . . . . . . 584.5 The exact values for lower-bounds. . . . . . . . . . . . . . . . . . . . . 614.6 The exact values for upper-bounds. . . . . . . . . . . . . . . . . . . . . 614.7 The approximate values for lower-bounds. . . . . . . . . . . . . . . . . 634.8 The approximate values for upper-bounds. . . . . . . . . . . . . . . . . 64
5.1 Approximate values for the lower-bounds. . . . . . . . . . . . . . . . . 755.2 Approximate values for the upper-bounds. . . . . . . . . . . . . . . . . 765.3 Approximate values for the lower-bounds. . . . . . . . . . . . . . . . . 825.4 Approximate values for the upper-bounds. . . . . . . . . . . . . . . . . 82
xxii
Dedicated to my son Kavya .........
xxiii
CHAPTER 1
General Introduction
1.1 Introduction
This chapter deals with the introductory nature and general introduction of the subject
matter of the thesis viz, the structure of the thesis, problem definition, objective and
future scope of work, original contribution by the thesis, methodology of research and
results, achievements with respect to objectives and conclusion.
Due to the vast scope of applications of the differential equations in the fields of
science, engineering, and technology, the specific problems are almost unlimited in a
different field that involves fuzziness and therefore it is reasonable to select such types
of problems for discussion here. Accordingly, the most interesting problems of current
interest have been used for mathematical treatment in the thesis. The explored problems
of the present study are related to the application of the solution of higher-order DEs that
contain fuzziness in initial and boundary conditions as well as in the forcing function.
Here, we are going to discuss fuzzy dynamic problems like the mass-spring system in
mechanical engineering, the bending of an elastic beam under the constant load in civil
engineering, the simple pendulum and Ariy’s non-homogeneous DE in physics(i.e.in
optics), etc. The mathematical problem of different physical phenomenon gives us a
one-dimensional DE with FIBCs. This Ph.D. work would be useful for solving fuzzy
dynamic problems arising in different fields of science, engineering, and technology.
1
General Introduction 2
1.2 Structure of the thesis
The thesis consists of 6 main chapters. After these introductory words, Chapter TWO
includes a brief discussion on certain relevant topics, like fuzzy set, membership
function, fuzzy number, fuzzy arithmetic, Zadeh’s extension principle, continuity,
Hukuhara difference, Hukuhara differential, generalized Hukuhara differentiability,
fuzzy Reimann integration and related literature review.
Chapter THREE deals with the solution of nth order DEs with constant coefficients
and Cauchy-Euler’s DEs with variable coefficients with fuzzy initial conditions(FICs)
by using properties of the linear transformation. We also solved one application-level
problem related to physics i.e. the study of motion of mass when the mass tapped
periodically with a hammer.
Chapter FOUR deals with the solution of nth order DEs with FIBCs by Gauss
elimination method(GEM) by using MATLAB and in comparing the exact and
approximate solutions. In this chapter, we solved a problem-related to civil
engineering - the bending of an elastic beam under the constant load.
Chapter FIVE deals with the solution of higher-order DEs with fuzzy boundary
conditions(FBCs) whose exact solution is difficult to find and interpret, that type of DEs
are solved by using the finite difference method(FDM). Here, we solve the differential
equation whose solution contains Airy’s function. The problems have been studied
numerically and graphically with the help of MATLAB.
Chapter SIX deals with the solution of some higher-order DEs with
constant-coefficients and variable coefficients (not a Cauchy-Euler’s) DEs with FIBCs
by using Laplace transform(LT). In this chapter, we solved problems related to the
mass-spring system in mechanical engineering branch and the simple pendulum
problem related to physics.
Some of the results/solutions reported here are also published in scholarly, peer-
reviewed and indexed journals as well as presented in the conference.
Brief description on the state of the art of the research topic 3
The references are given alphabetically at the end.
1.3 Brief description on the state of the art of the
research topic
Lotfi Zadeh introduced the concept of fuzzy set theory and pointed out ambiguities and
impreciseness in his seminal paper entitled "Fuzzy Sets" [80] in 1965. Here, we are
going to study the dynamic problems with fuzzy parameters which occurs in different
fields like mechanical engineering, civil engineering, physics, etc. All these branches
of sciences and engineering have contributed to a vast amount of literature on this topic.
The notion of a fuzzy derivative was first introduced by Chang and Zadeh [25], it
was reviewed by Dubois and Prade [31] who used the extension principle [79] in their
approach. Other fuzzy derivative concepts were introduced by Puri and Ralescu [69]
and Goetschel and Voxman [40] as an extension of HD for multivalued functions.
Kandel and Byatt [48] used the concept of FDE for the analysis of fuzzy dynamical
problems. The FDEs were discussed by Kaleva in [44, 46] and "The Cauchy problem
for fuzzy differential equations" was studied by Kaleva in [45]. Seikkala [74]
discussed "On the fuzzy initial value problems". Buckley and Feuring [22] approached
to "Fuzzy initial value problem for nth order linear differential equations". "The
Cauchy problem for continuous fuzzy differential equations" was discussed by Nieto
[59]. Mondal et al. [58] presented the solution technique for the "First order linear non
homogeneous ordinary differential equation in fuzzy environment based on Laplace
transform". "Existence of the solutions of fuzzy differential equations with
parameters" was discussed by Ding et al. [30]. Existence and uniqueness of solution to
FBVP have been proved by Esfahani et al. [32]. "Existence and uniqueness of
solutions to Cauchy problem of fuzzy differential equations" have been discussed by
Song et al. [75]. Lakshikantham et al. [55] worked on "Interconnection between set
and FDEs" and on the solution of "Two-point boundary value problem associated with
the non-linear fuzzy differential equation" [53].
General Introduction 4
The idea of GHD was employed by Bede and Gal [18, 20] to investigate
first-order linear FDEs. The "Remark on the new solution of FDEs" was given by Bede
and Gal [19]. GHD was recommended by Chalco-Cano and Romon Flores [24] for
solving FDEs. "First order linear FDEs under generalized differentiability" was
discussed by Bede et al. [21]. Allahviranloo et al. [10] found solution of "nth order
fuzzy linear differential equations". Allahviranlloo et al. [8] obtained "A new method
for solving fuzzy linear differential equations". The existence and uniqueness of
solutions of second-order FDEs has proposed by Allahviranloo et al. [12]. The
existence and uniqueness theorem of the solution of nth order FDE and the integral
form of nth order FDE were studied by Salahshour [71] under GHD. Salahshour and
Allahviranloo [72] worked on "Application of fuzzy Laplace Transform". Xiaobin et
al. [78] found "Fuzzy approximate solutions of second-order fuzzy linear boundary
value problems". "A boundary value problem for second-order fuzzy differential
equations" was studied by Khastan et al. [50]. "A new result on multiple solutions for
nth order fuzzy differential equation under generalized differentiability" has been
suggested by Bahrami et al. [15]. Ahmad et al. [4] studied analytic and numerical
solution of the FDEs based on the extension principle. Ivaz et al. [42] investigated "A
numerical algorithm for the solution of first-order fuzzy differential equations and
hybrid fuzzy differential equations". "Variation of constant formula for first-order
FDEs" has been presented by Khastan et al. [49]. Akin et al. [6] developed "An
algorithm for the solution of second-order FIVPs".
Mansouri et al. [57] offered "A numerical method for solving nth order FDEs by
using characterization theorem". "Numerical solutions of fuzzy differential and
integral equations" were proposed by Friedman et al. [33]. "A new fuzzy version of
Euler’s method for solving differential equations with fuzzy initial values" were
introduced by Ahmad and Hasan [3]. "Homotopy perturbation method for solving nth
order fuzzy linear differential equations" was implemented by Tapaswini and
Chakraverty [76]. "The variational iteration method for solving nth order FDEs." was
discussed by Jafari et al. [43]. The different methods for solving nth order FDEs were
discussed in [5, 10, 11, 15, 66, 71].
Definition of the problem 5
1.4 Definition of the problem
In the present study different dynamic problems with FIBCs that occur in many fields
of engineering, science and technology are discussed. Here, we discuss dynamic
problems with fuzzy parameters related to mechanical engineering, civil engineering,
and physics, etc by using a different method and compare the exact and approximate
solutions by using graphical representation. The aim of the present study was to
investigate the behaviour of the solution of the DEs with FIBCs which are arising in
the different fields and also approximated the solution of the Airy’s non-homogeneous
DE by considering fuzzy boundaries. The objective of the work is to investigate the
behaviour of the solution of higher-order ordinary differential equations with fuzzy
parameters and study the fuzziness involved.
1.5 Objective and future scope of work
The main goal of the present work is to study the DEs with fuzzy parameters that are
arising in many fields of engineering and sciences. The objectives of our study are as
under:
• To find the fuzziness in the solution to dynamic problems due to the fuzziness in
the initial conditions and the forcing function,
• To find the fuzziness in the solution to dynamic problems due to the fuzziness in
the boundary conditions and the forcing function,
• To find the fuzziness in the solution to dynamic problems due to the fuzziness in
the initial and the boundary conditions as well as the forcing function.
Different methods are used to study the effect of FIBCs in the final solution of DEs. The
future scope of the current work is to study the problems related to partial differential
equations with FIBCs as well as fuzziness in the forcing function.
General Introduction 6
1.6 Original contribution by the thesis
The original contribution made by the study is the solution of nth order DE with
constant coefficients and Cauchy-Euler’s with variable coefficients with FICs by using
the properties of a linear transformations. The GEM is used to solve nth order DEs
with FIBCs using MATLAB, and the exact and approximate solutions are compared.
In chapter FOUR, we have solved a problem related to the civil engineering viz., the
bending of an elastic beam under the constant load.
The thesis includes the solution of higher-order DEs with FBCs whose exact
solution is difficult to find and interpret that type of DEs are solved by FDM. Here, we
also solved the differential equation that contains the Airy’s function. The solutions of
the problems has been studied numerically and graphically with the help of MATLAB.
The solutions of some constant-coefficient and variable coefficient(not a
Cauchy-Euler’s) DEs with FIBCs is obtained by using Laplace transform(LT). In last
chapter, we solved a problem-related to mechanical engineering viz., the mass-spring
system and problem-related to physics viz., the simple pendulum. The solution of
higher-order DEs with FIBCs that occur in different fields of engineering and sciences
are discussed in this thesis.
1.7 Methodology of Research and Results/Comparisons
We have studied various kinds of literature related to the solution of different types of
higher-order DEs with FIBCs and did a comparative analysis to find the research gap
and the problem statement. The literature survey helped us to define an objective of the
research.
We have used MATLAB coding to solve various DEs with fuzzy parameters. We
have also used Mathematica and MATLAB software for graphical representation of
different types of solution bounds. It is a combination of different solution methods to
solve DEs with FIBCs that provides a convenient analytic and numerical tool to solve
Achievements with respect to objectives 7
many types of ordinary differential equations. We have studied various problems
related to engineering and science and according to research gap, we have created the
mathematical models for different physical phenomenon arising in the fields of
mechanical engineering, civil engineering, and physics.
An approximate analytical solution of nth order ordinary differential
equations(ODEs) with FICs and the study of the motion of mass arising in physics are
obtained by the properties of the linear transformation. The GEM has been used to find
the solution of nth order DEs with appropriate FIBCs. The solution is interpreted
numerically and graphically using MATLAB and Mathematica software and it is
compared with the exact solution. The FDM is used to approximate the solution of
higher-order DEs with FBCs and also solved the Airy’s non-homogeneous DE by the
help of MATLAB coding and plotting. The variable coefficient(not a Cauchy-Euler’s)
DEs with FIBCs are solved by using the method of LT. Here, we solved the simple
pendulum problem in physics and the mass-spring problem in mechanical engineering
branch by using LT. The solution is interpreted numerically and graphically by using
Mathematica software and it is compared with the exact solution.
1.8 Achievements with respect to objectives
A mathematical model is developed for the problem related to the mass-spring system in
the mechanical engineering with FICs, the bending of an elastic beam under the constant
load in civil engineering with FIBCs, the second-order linear differential equation viz.,
Airy’s non-homogeneous DE with FBCs, and the simple pendulum problem in physics
are solved by various methods.
Here, we discuss different solution methods to solve different type of DEs with
FIBCs as well as fuzziness in the forcing function. We find the solution of nth order
DEs with FICs by using properties of linear transformation and discuss the application
related to the study of the motion of the mass in physics. The solution of nth order DEs
with FIBCs is obtained using the GEM and an application related to civil engineering
viz., the bending of an elastic beam under the constant load is studied. We also find the
General Introduction 8
solution of higher-order DEs with FBCs by using the FDM and solved the Airy’s
non-homogeneous DE with FBCs. In last, we have solved higher-order variable
coefficient(not a Cauchy-Euler’s) DEs with FIBCs by using LT method and the
mass-spring system and the simple pendulum problem in mechanical engineering and
physics respectively.
1.9 Conclusion
We have studied different methods of solving DEs with FIBCs arising in different filed
of engineering such as mechanical, civil, and the Airy’s non-homogeneous DE and the
simple pendulum problem in physics. Here, we have discussed the analytic and the
numerical methods to solve different types of DEs with FIBCs. We have solved DEs
with FIBCs whose exact solution is known then we compare the exact and
approximate solutions of higher-order DEs. We have also approximated the solution of
DEs with FBCs whose exact solution is not known, the solution of such type of DEs is
approximated by numerical method i.e. using FDM. The mass-spring problem in
mechanical engineering and the simple pendulum problem in physics were solved
using the Laplace transform.
CHAPTER 2
Fundamentals of Fuzzy Set Theory and
Basic Concepts
2.1 History of development of fuzzy set theory
Lotfi Zadeh was best known for proposing fuzzy Mathematics. Fuzzy set theory is a
process to deal with imprecise data. Fuzzy logic is a form of many-valued logic; it deals
with reasoning that is approximate rather than fixed and exact. Compared to traditional
binary sets(true or false values) fuzzy logic variables may have truth value that ranges
in degree between 0 and 1(no and yes).[https://www.igi-global.com/dictionary/fuzzy-
logic/11740]
FIGURE 2.1: Lotfi Zadeh [https://www.slideserve.com/tambre/lotfi-zadeh]
9
Fundamentals of Fuzzy Set Theory and Basic Concepts 10
2.2 Crisp set vs. Fuzzy set
Let U be universe of discourse or a universal set.
2.2.1 Crisp set
Crisp set A is defined by its characteristic function. The characteristic function is
denoted by χA and it is defined by
χA(x) =
1, if x ∈ A
0, if x 6∈ A
It takes value either 1 or 0.
2.2.2 Fuzzy set
Let U 6= φ . The fuzzy set A ∈U is defined by its membership function.
µA : U → [0,1]
It takes values in the range [0,1].
U
x
µA(x)
Fuzziness Fuzziness
1
o
Crisp Subset A
Fuzzy Subset A
FIGURE 2.2: Crisp set vs. Fuzzy set
Standard operations on fuzzy sets: 11
2.3 Standard operations on fuzzy sets:
The conjunction, dejection and complement are three standard properties for fuzzy sets.
These are true for both fuzzy set and crisp set. The fuzzy set theory defines fuzzy
operations on fuzzy sets. The AND, OR, NOT of Boolean logic exist in fuzzy logic
usually defined as minimum, maximum, and complement, respectively [70].
2.3.1 Logical connectives
Let A and B are fuzzy sets
• Conjunction or Intersection:
A∧ B = A∩ B = min(µA(x),µB(x))
• Disjunction or Union:
A∨ B = A∪ B = max(µA(x),µB(x))
• Complementary:
A→ µA′ (x) = 1−µA(x)
x
µA
5 8
1
0
A
FIGURE 2.3: A
Fundamentals of Fuzzy Set Theory and Basic Concepts 12
x
µB
4
1
0
B
FIGURE 2.4: B
x
µA∧B
4
1
0 5 8
1B A
FIGURE 2.5: A∧ B
x
µA∨B
4
1
0 5 8
1B A
FIGURE 2.6: A∨ B
x
µA′
5 8
1
0
A
FIGURE 2.7: A′
Fuzzy set and Membership function 13
The below two laws do not hold for fuzzy sets because of the fact that, since fuzzy sets
can overlap, a set and its complement can also overlap.
Axiom of the contradiction A∧ A′= A∩ A
′ 6= φ
Axiom of the excluded-middle A∨ A′= A∪ A
′ 6=U
2.4 Fuzzy set and Membership function
The membership function (MF) [70] incorporate the mathematical representation of
membership in a set, and the notation used throughout the thesis for a fuzzy set is a
set symbol with a tilde over-score, say A, where the functional mapping is given by
µA(x) ∈ [0,1] and the symbol µA is the degree of membership of element x in fuzzy set
A. Therefore, µA(x) is a value on the unit interval that measures the degree to which
element x belongs to fuzzy set A ; equivalently,
µA(x) = degree to which x ∈ A.
2.4.1 Membership function
A fuzzy set A can be defined as a pair of the universal set U and MF µA : U → [0,1] for
each x ∈ U , the number µA is called the membership degree of x in A. It can be
represented mathematically as
A = {(x,µA(x))|x ∈U}, µA(x) ∈ [0,1].
2.5 Properties of Membership function
All the information contained in a fuzzy set is described by its membership function.
Fundamentals of Fuzzy Set Theory and Basic Concepts 14
2.5.1 Core
The core [70] of a MF for some fuzzy set A is defined as the region of the universe U
that is characterized by complete and full membership in the set A. That is, the core
comprises those elements x of the universe such that µA(x) = 1.
2.5.2 Support
The support [70] of a MF for some fuzzy set A is defined as the region of the universe
U that is characterized by a non-zero membership in the set A. That is, the support
comprises those elements x of the universe such that µA(x) > 0. The support of A is a
crisp set and it is defined as
supp (A) = {x ∈U |µA(x)> 0}
2.5.3 Boundary
The boundaries of a MF for some fuzzy set A is defined as the region of the universe U
that have a non-zero membership but not a complete membership. That is, the
boundaries contains those elements x of the universe such that
0 < µA(x)< 1.
These are the elements of the universe that have some degree of fuzziness or only
partial membership in the fuzzy set A. Fig.2.8 [70] shows the regions in the universe
comprising the core, support, and boundaries of a typical fuzzy set.
Properties of Membership function 15
x
µA(x)
Core
Support
Boundary Boundary
1
o
FIGURE 2.8: Core, support, and boundaries of a fuzzy set
2.5.4 Height
The height [70] of a fuzzy set A is the largest membership grade of any element in A.
h(A) = height(A) = max {µA(x),x ∈ A}.
2.5.5 Normal fuzzy set
A normal fuzzy [70] set is the one whose MF has at least one element x in the universe
whose membership value is unity. i.e. h(A) = 1
x
µA(x)
1A
0
FIGURE 2.9: Normal fuzzy set
Fundamentals of Fuzzy Set Theory and Basic Concepts 16
2.5.6 Subnormal fuzzy set
If h(A)< 1,∀x ∈U , then the fuzzy set is called a subnormal fuzzy set [70].
x
µA(x)
1
A
0
FIGURE 2.10: Subnormal fuzzy set
2.5.7 Convex fuzzy set
A convex fuzzy set is described by a MF whose membership values are strictly
monotonically increasing, or whose membership values are strictly monotonically
decreasing, or whose membership values are strictly monotonically increasing then
strictly monotonically decreasing with increasing values for elements in the universe.
In notations, if for any elements x,y and z in a fuzzy set A,
x < y < z⇒ µA(y)≥ min[µA(x),µA(z)]
then A is said to be a convex fuzzy set [70].
Fuzzy number 17
x
µA(x)
1
0 x zy
A
FIGURE 2.11: Convex fuzzy set
2.5.8 r-cut
It is one of the most important concepts of fuzzy sets.
The r-cut [70] of A is a crisp set and it is denoted as Ar and defined as
Ar = {x ∈U |µA(x)≥ r}
The strong r-cut [70] of A is also a crisp set and it is denoted as Ar+ and defined as
Ar+ = {x ∈U |µA(x)> r}
For r = 0, A0 = closure(supp(A))
2.6 Fuzzy number
A fuzzy number is a fuzzy set defined on the set of real numbers IR like µ : IR→ [0,1].
which satisfies:
• µ is upper semi-continuous i.e. ∀ ε > 0, ∃ δ > 0, 3 µ(x)−µ(x0)< ε whenever
|x− x0|< δ ,
Fundamentals of Fuzzy Set Theory and Basic Concepts 18
• µ is fuzzy convex i.e µ(λx+(1−λ )y)≥min{µ(x),µ(y)} ∀ x,y∈ IR, λ ∈ [0,1],
• µ is normal i.e ∃ x0 ∈ IR for which µ(x0) = 1,
• supp µ = {x ∈ IR | µ(x) > 0} is support of µ , and its closure cl(supp µ) is
compact.
Let E be set of all real fuzzy numbers [44] which are normal, upper semi-continuous,
convex and compactly supported fuzzy sets. The parametric form a fuzzy number µ is
a order pair (µ,µ) of functions (µ(r),µ(r)),0 ≤ r ≤ 1, which satisfies the following
requirements:
• µ(r) is a bounded monotonic increasing left continuous function over [0,1];
• µ(r) is a bounded monotonic decreasing right continuous function over [0,1];
• µ(r)≤ µ(r),0≤ r ≤ 1.
2.6.1 Triangular fuzzy number
Consider triangular fuzzy number (TFN) [76] A = (a,b,c) as seen in Fig.2.12. The
TMF µA of A is defined as follows.
µA(x) =
0 , if x < ax−ab−a , if a≤ x≤ bc−xc−b , if b≤ x≤ c
0 , if x > c
The TFN A = (a,b,c) can be written as an order pair of function of r-cut approach
i.e. [µ(r), µ(r)] = [a+(b−a)r, c− (c−b)r], where r ∈ [0,1].
Fuzzy number 19
x
µA(x)
a cb
1
0
FIGURE 2.12: Triangular fuzzy number
2.6.2 Trapezoidal fuzzy number
Consider trapezoidal fuzzy number (TraFN) [82] A = (a,b,c,d) as presented in
Fig.2.13. The TraMF µA of A is defined as follows.
µA(x) =
0 , if x < ax−ab−a , if a≤ x≤ b
1 , if b≤ x≤ cd−xd−c , if c≤ x≤ d
0 , if x≥ d
The TraFN A = (a,b,c,d) can be written as an order pair of function of r-cut approach
i.e. [µ(r), µ(r)] = [a+(b−a)r, d− (d− c)r], where r ∈ [0,1]
Fundamentals of Fuzzy Set Theory and Basic Concepts 20
x
µA(x)
a cb d
1
o
FIGURE 2.13: Trapezoidal fuzzy number
2.6.3 Gaussian fuzzy number
The Gaussian membership function (GMF) [82] µA of an asymmetric Gaussian fuzzy
number(GFN) A = (α,σl,σr) is defined as follows.
µA(x) =
e− (x−α)2
2σ2l , if x≤ α
e− (x−α)2
2σ2r , if x≥ α
where, α denotes the modal value and σl,σr denote left and right fuzziness,
respectively, corresponding to the Gaussian distribution.
For symmetric GFN the left and right fuzziness are equal i.e. σl = σr = σ . So, the
symmetric GFN is denoted as A = (α,σ ,σ) and corresponding GMF is defined as
µA(x) = e−β (x−α)2,∀x ∈ IR where β = 1
2σ2 . The symmetric GFN in parametric form
can be written as A = [µ(r),µ(r)] =[α−
√− (loger)
β, α +
√− (loger)
β
]where r ∈ [0,1].
Fuzzy arithmetic 21
0.368
x0
1
µA(x)
FIGURE 2.14: Gaussian fuzzy number
2.7 Fuzzy arithmetic
Let u =(u(r),u(r)),v = (v(r), v(r)) ∈ E, r ∈ [0,1] and arbitrary k ∈ IR [76].
Then
u = v iff u(r) = v(r) and u(r) = v(r),
u+ v = (u(r)+ v(r), u(r)+ v(r)),
u− v = (u(r)− v(r), u(r)− v(r)),
ku =
(ku(r),ku(r)), k ≥ 0
(ku(r),ku(r)), k < 0
u · v = (min{u(r)v(r), u(r)v(r), u(r)v(r), u(r)v(r)},max{u(r)v(r), u(r)v(r), u(r)v(r), u(r)v(r)})
2.8 Zadeh’s extension principle
Zadeh’s extension principle [79] provides a mathematical approach for extending
classical functions to fuzzy mappings. Let X be a set, P(X) be the power set of X , and
Fundamentals of Fuzzy Set Theory and Basic Concepts 22
F(X) be fuzzy power set of X . That is,
P(X) = {U |U ⊂ X}= {U |U : X →{0,1}}
and
F(X) = {U |U : X → [0,1]}
Let f be a mapping from a set X to another set Y and fmax be mapping from F(X) to
F(Y ) which can be induced by f as following:
fmax : F(X)→ F(Y ),u→ fmax(u)
where ∀y ∈ Y ,
fmax(u)(y) =
∨f (x)=y u(x), i f
{x| f (x) = y,x ∈ X
}6= φ
0, otherwise
2.9 Continuity
Let f : IR→ E, f is called continuous [33], if for every t0 ∈ (a,b) and every ε > 0,
there exists δ > 0 such that if |t− t0|< δ then D( f (t), f (t0))< ε .
A mapping f : IR×E→ E is called continuous [75] at point (t0,x0) ∈ IR×E provided
for any fixed r ∈ [0,1] and arbitrary ε > 0, there exist a δ (ε,r) such that
D([ f (t,x)]r, [ f (t0,x0)]r)< ε whenever | t− t0 |< δ (ε,r).
2.10 Hukuhara difference
Let u,v ∈ E. If there exists w ∈ E such that u = v+w, then w is called the Hakuhara-
difference [18] of fuzzy numbers u and v, and it is denoted by w = u v.
The sign stands for Hakuhara-difference, and u v 6= u+(−1)v.
Hukuhara differential 23
2.11 Hukuhara differential
Let f : (a,b)→ E and t0 ∈ (a,b). We say that f is Hukuhara differentiable(HD) [18] at
t0, if there exists an element f′(t0) ∈ E such that for all h > 0 sufficiently small,
∃ f (t0 +h) f (t0), f (t0) f (t0−h) and the limits holds (in the metric D)
limh→0
f (t0 +h) f (t0)h
= limh→0
f (t0) f (t0−h)h
= f′(t0).
2.12 Generalized Hukuhara differentiability
Let f : (a,b)→ E and t0 ∈ (a,b). We say that f is (1)-differential at t0, if there exists
an element f′(t0) ∈ E such that for all h > 0 sufficiently small,
∃ f (t0 +h) f (t0), f (t0) f (t0−h) and the limits holds (in the metric D)
limh→0
f (t0 +h) f (t0)h
= limh→0
f (t0) f (t0−h)h
= f′(t0).
and f is (2)-differentiable if for all h > 0 sufficiently small,
∃ f (t0) f (t0 +h), f (t0−h) f (t0) and the limits holds (in the metric D)
limh→0
f (t0) f (t0 +h)−h
= limh→0
f (t0−h) f (t0)−h
= f′(t0).
If f′(t0) exist in above two cases then i.e called generalized Hukuhara
derivative(GHD)[24] of f (t).
2.13 Fuzzy Reimann integration
Let f (t) be fuzzy-valued function on [0,∞) and it is represented by ( f (t,r), f (t,r)).
For any fixed r ∈ [0,1], assume f (t,r) and f (t,r) are Reimann-integrable [77], on [a,b]
for every b ≥ a and assume there are two positive M(r) and M(r) such that
Fundamentals of Fuzzy Set Theory and Basic Concepts 24
∫ ba | f (t,r) | dt ≤ M(r) and
∫ ba | f (t,r) | dt ≤ M(r) for every b ≥ a, then f (t) is
improper fuzzy Reimann integral(FRI) on [a,∞) and the improper FRI is a fuzzy
number. Furthermore, we have
∫∞
a f (t)dt =(∫
∞
a f (t,r)dt,∫
∞
a f (t,r)dt)
CHAPTER 3
Solutions of nth order Differential
Equations with Fuzzy Initial
Conditions by using the Properties of
Linear Transformations
3.1 Introduction
These days, the DEs with FICs is a prominent theme examined by numerous specialists
since it is widely used for the purpose of modelling in science and engineering. The
DEs with FICs emerge normally during the modelling of dynamic problems. The FDE
was studied by Kaleva [44, 46]. FIVPs were studied by Seikkala [74]. In this chapter,
the derivatives are considered as Hukuhara derivatives.
Bede et al. [17] used the notions of Hukuhara differential, generalized
differentiability, differential inclusions to solve FDEs and did the interpretation of
FDEs by using Zadeh’s extension principle on the classical solution. A new approach
to fuzzy initial value problem was studied by Gasilov et al. [37]. They proposed a new
approach to non-homogeneous fuzzy initial value problem [38]. Georgiou et al. [39]
solved IVPs for higher-order FDEs. Analytical solutions of FIVPs by homotopy
analysis method were studied by Abu-Arqub et al. [1]. Solution of linear differential
equations with fuzzy boundary values was put forwarded by Gasilov et al. [34, 35].
Gasilov et al. [36] also found solution method for a BVP with fuzzy forcing function.
25
Solutions of nth order Differential Equations with Fuzzy Initial Conditions by using theProperties of Linear Transformations 26
3.2 Linear transformation
Linear transformation is rule that converts one data into another. It is a function (say T )
that takes each element of the domain and associate it with exactly one element of the
co-domain.
3.2.1 Wronskain
If the functions f1 = f1(x), f2 = f2(x), ...., fn = fn(x) have n−1 continuous derivatives
on the interval (−∞,∞), and if the Wronskian of these functions
W (x) =
∣∣∣∣∣∣∣∣∣∣∣∣
f1(x) f2(x) . . . fn(x)
f1′(x) f2
′(x) . . . fn
′(x)
...... . . . ...
f1n−1(x) f2
n−1(x) . . . fnn−1(x)
∣∣∣∣∣∣∣∣∣∣∣∣
is not identically zero on (−∞,∞), then these functions form a linearly independent set
of vectors in C(n−1)(−∞,∞).
3.3 Problem formulation
In real life practical problems, the behaviour of an entity is obtained with the help of
physics and it is crisp. But the initial conditions are obtained from calculations, these
values can be vague and often it is convenient to model them using different types of
fuzzy numbers. This gives rise to IVPs with crisp dynamics but fuzzy conditions.
In the field of DEs, an IVP is an ODE together with a designated value, called
the initial condition, of the unknown function at a prescribed point in the domain of
the solution. Here we are going to solve nth order DEs by considering fuzzy initial
conditions. For same value of the independent variable, there should be a range of values
of the dependent variable since we are considering values of the dependent variable i.e.
initial conditions to be fuzzy. We show that if the solution of the corresponding crisp
Method of solving nth order differential equation with fuzzy initial conditions 27
problem exists and is unique then the fuzzy problem also has a unique solution. Here
we are going to solve homogeneous and non-homogeneous constant coefficient as well
as Cauchy-Euler’s variable coefficient DEs with FICs.
3.4 Method of solving nth order differential equation
with fuzzy initial conditions
The notation u = (uL(r), uR(r)),0 ≤ r ≤ 1 will be used to represent fuzzy number in
parametric form.
Let u = uL(0) and u = uR(0) are used to indicate the left and the right end points of u
respectively.
An r-cut of u is an interval [uL(r), uR(r)], which is denoted as
ur = [ur, ur]
Consider the triangular fuzzy number u = (a,c,b) for which
uL(r) = a+ r(c−a), uR(r) = b− r(b− c)
Its geometric interpretation is u = a,u = b are end points and c is the vertex. Let us
consider the TFN whose vertex is zero i.e. u = (a,0,b) (note that a < 0 and b > 0).
Then,
u = [uL(r), uR(r)] = [(1− r)a, (1− r)b] = (1− r)[a,b], r ∈ [0,1].
From the above result, we can say that r-cut is similar to interval [a,b] (i.e. to 0-cut)
with similarity coefficient (1− r).
A fuzzy number u can be express as
u = ucr + uun
Solutions of nth order Differential Equations with Fuzzy Initial Conditions by using theProperties of Linear Transformations 28
Here ucr is a fuzzy number with membership degree 1 and represent the crisp part(i.e.
vertex) of u; whereas uun is uncertain part with vertex at origin.
For triangular fuzzy number u = (a,c,b)
ucr = c and uun = (a− c,0,b− c).
Consider the nth order DE with FICs:
bn(t)xn +bn−1(t)xn−1 + ....+b1(t)x′+b0(t)x = f (t)
x(l) = A1
x′(l) = A2
... =...
xn−1(l) = An
(3.1)
where A1, A2, ....., An are fuzzy numbers and b0(t),b1(t),b2(t), ....,bn(t) and f (t) are
continuous crisp functions or constants and l is any integer.
Let us represent the initial values as
A1 = a1 + a1
A2 = a2 + a2
... =...
...
An = an + an
where a1,a2, ....an are crisp numbers, while a1, a2, ..., an are fuzzy numbers. Let us
split the DE with FICs in (3.1) to following problems:
The homogeneous DE with FICs
bn(t)xn +bn−1(t)xn−1 + ....+b1(t)x′+b0(t)x = 0
x(l) = a1
x′(l) = a2
... =...
xn−1(l) = an
(3.2)
Method of solving nth order differential equation with fuzzy initial conditions 29
and a non-homogeneous DE with crisp conditions
bn(t)xn +bn−1(t)xn−1 + ....+b1(t)x′+b0(t)x = f (t)
x(l) = a1
x′(l) = a2
... =...
xn−1(l) = an
(3.3)
It is easy to see that if xun(t) and xcr(t) are solutions of (3.2) and (3.3) respectively, then
x(t) = xcr(t)+ xun(t) is a solution of given problem (3.1).
To determine x(t) we consider the transformation T : IRn→ IR,T (u) = v · u is a linear
transformation,
where v is fixed n×n determinant and u = [a1,a2, . . . ,an]T .
The solution algorithm consist four steps:
1. Represent the initial values as A1 = a1 + a1, A2 = a2 + a2,..., An = an + an.
2. Find linearly independent solutions x1(t),x2(t), ...,xn(t) of the crisp differential
equation bn(t)xn +bn−1(t)xn−1 + ....+b1(t)x′+b0(t)x = 0. Constitute the vector
function s(t) = (x1(t),x2(t), ....,xn(t)), the determinant W by using the definition
of Wronskain and calculate the vector function
y(t) = s(t)W−1 = (y1(t),y2(t), ....,yn(t)).
3. Find solution xcr(t) of the non-homogeneous crisp problem.
4. Find the solutions of the given problems
(a) For homogeneous DE with FICs
x(t) = y1(t)a1 + y2(t)a2 + ......+ yn(t)an
(b) For non-homogeneous DE with FICs
x(t) = xcr(t)+ y1(t)a1 + y2(t)a2 + ......+ yn(t)an
Solutions of nth order Differential Equations with Fuzzy Initial Conditions by using theProperties of Linear Transformations 30
3.5 Solution of a higher order differential equation with
fuzzy initial condition
3.5.1 Solution of a higher order constant coefficient differential
equation with fuzzy initial condition
Example 3.1 Solve the constant coefficient homogeneous DE with FICs
x′′+4x
′+3x = 0,
x(0) = (−0.5,0,0.5)
x′(0) = (−1,0,0.5)
The problem is homogeneous and initial values are fuzzy numbers with vertices at 0.
Therefore, the solution using the algorithm is obtained as follows:
x1(t) = e−t and x2(t) = e−3t are linearly independent solutions for the equation
x′′+4x
′+3x = 0. Hence s(t) = (e−t ,e−3t) and
W =
∣∣∣∣∣∣1 1
−1 −3
∣∣∣∣∣∣
and
y(t) = (y1(t),y2(t))
=
(32
e−t− 12
e−3t ,12
e−t− 12
e−3t).
xun(t) =(
32
e−t− 12
e−3t)(−0.5,0,0.5)+
(12
e−t− 12
e−3t)(−1,0,0.5),
where arithmetic operations are considered to be fuzzy operations. The fuzzy solution
x(t) form band in the tx-coordinate space (Fig.3.1).
Since the initial values are TFNs, an r-cut of the solution can be determined by similarity
coefficient (1− r), i.e. xr(t) = (1− r)[ x(t), x(t) ].
Solution of a higher order differential equation with fuzzy initial condition 31
FIGURE 3.1: The fuzzy solution obtained by the properties of linear transformations
Example 3.2 Consider the constant coefficient non-homogeneous DE with FICs
x′′+4x
′+3x = e−t ,
x(0) = (0.5,1,1.5)
x′(0) = (1,2,2.5)
We represent the FICs as
A = (0.5,1,1.5) = 1+(−0.5,0,0.5),
B = (1,2,2.5) = 2+(−1,0,0.5).
The crisp solution to the crisp non-homogeneous problem
x′′+4x
′+3x = e−t ,
x(0) = 1
x′(0) = 2
is
xcr(t) =94
e−t− 54
e−3t +t2
e−t
=
(32
e−t− 12
e−3t)+
32
(e−t
2− e−3t
2
)+
t2
e−t
Solutions of nth order Differential Equations with Fuzzy Initial Conditions by using theProperties of Linear Transformations 32
The fuzzy homogeneous problem to find the uncertainty of the solution is as follows:
x′′+4x
′+3x = 0,
x(0) = (−0.5,0,0.5)
x′(0) = (−1,0,0.5)
This problem is the same as Example 3.1 Hence, its solution is
xun(t) =(
32
e−t− 12
e−3t)(−0.5,0,0.5)+
(12
e−t− 12
e−3t)(−1,0,0.5)
Adding this uncertainty to the crisp solution gives the fuzzy solution of the given DE
with FICs
x(t) = xcr(t)+ xun(t)
=
(32
e−t− 12
e−3t)(0.5,1,1.5)+
(12
e−t− 12
e−3t)(1,2,2.5)+
t2
e−t
The fuzzy solution x(t) form band in the tx-coordinate space (Fig.3.2).
FIGURE 3.2: The fuzzy solution obtained by the properties of linear transformationsand the red line represents the crisp solution
Solution of a higher order differential equation with fuzzy initial condition 33
Example 3.3 Solve the constant coefficient homogeneous DE with FICs.
x′′′+3x
′′+3x
′+ x = 0,
x(0) = (−0.5,0,1)
x′(0) = (−1,0,1)
x′′(0) = (−1,0,0.5)
The problem is homogeneous and initial values are fuzzy numbers with vertices at 0.
Therefore, the solution is by solution algorithm.
x1(t) = e−t , x2(t) = te−t and x3(t) = t2e−t are linearly independent solutions for the
equation x′′′+3x
′′+3x
′+ x = 0. Hence s(t) = (e−t , te−t , t2e−t) and
W =
∣∣∣∣∣∣∣∣∣
1 0 0
−1 1 0
1 −2 2
∣∣∣∣∣∣∣∣∣
and
y(t) = (y1(t),y2(t),y3(t))
=(
e−t + te−t +t2
2e−t , te−t + t2e−t ,
t2
2e−t)
xun(t) =(
e−t +te−t +t2
2e−t)(−0.5,0,1)+(te−t +t2e−t)(−1,0,1)+
t2
2e−t(−1,0,0.5);
where arithmetic operations are considered to be fuzzy operations.
The fuzzy solution x(t) form band in the tx-coordinate space (Fig.3.3).
Since the initial values are TFNs, an r-cut of the solution can be determined by similarity
coefficient (1− r), i.e. xr(t) = (1− r)[ x(t), x(t) ].
Solutions of nth order Differential Equations with Fuzzy Initial Conditions by using theProperties of Linear Transformations 34
FIGURE 3.3: The fuzzy solution obtained by the properties of linear transformation
Example 3.4 Consider the constant coefficient non-homogeneous DE with FICs.
x′′′+3x
′′+3x
′+ x = 30e−t ,
x(0) = (2.5,3,4)
x′(0) = (−4,−3,−2)
x′′(0) = (−48,−47,−46.5)
We represent the FICs as
A = (2.5,3,4) = 3+(−0.5,0,1),
B = (−4,−3,−2) =−3+(−1,0,1),
C = (−48,−47,−46.5) =−47+(−1,0,0.5)
We solve crisp non-homogeneous problem
x′′′+3x
′′+3x
′+ x = 30e−t ,
x(0) = 3
x′(0) =−3
x′′(0) =−47
Solution of a higher order differential equation with fuzzy initial condition 35
And the crisp solution is
xcr(t) = (3−25t2)e−t +5t3e−t
= 3(
e−t + te−t +t2
2e−t)−3(te−t + t2e−t)− 47
2t2e−t +5t3e−t
The fuzzy homogeneous problem to find the uncertainty of the solution is as follows:
x′′′+3x
′′+3x
′+ x = 0,
x(0) = (−0.5,0,1)
x′(0) = (−1,0,1)
x′′(0) = (−1,0,0.5)
This problem is the same as Example.3.3. Hence, its solution is
xun(t) =(
e−t + te−t +t2
2e−t)(−0.5,0,1)+(te−t + t2e−t)(−1,0,1)+
t2
2e−t(−1,0,0.5)
Adding this uncertainty to the crisp solution gives the fuzzy solution of the given DE.
x(t) = xcr(t)+ xun(t)
x(t) =(
e−t + te−t +t2
2e−t)(2.5,3,4)+(te−t + t2e−t)(−4,−3,−2)
+t2
2e−t(−48,−47,−46.5)+5t3e−t
The fuzzy solution x(t) form band in the tx-coordinate space (Fig.3.4).
Solutions of nth order Differential Equations with Fuzzy Initial Conditions by using theProperties of Linear Transformations 36
FIGURE 3.4: The fuzzy solution obtained by the properties of linear transformations,and the red line represents the crisp solution
In above example, if we take t → ∞ then the fuzziness in the solution is disappearing.
So, the solution goes nearer to zero if we increase t (Fig.3.5).
FIGURE 3.5: If t increases, fuzziness disappears
Solution of a higher order differential equation with fuzzy initial condition 37
3.5.2 Solution of higher order Cauchy-Euler’s variable coefficient
differential equations with fuzzy initial conditions
Example 3.5 Consider the variable coefficient DE with FICs
(t2D2−3tD+4)x = t2
x(1) = (0.5,1,2)
x′(1) = (−0.5,0,0.5)
(t2D2−3tD+4)x = t2
x(1) = 1
x′(1) = 0
Associated non-homogeneous problem has the solution
xcr = [1−2log(t)]t2 +t2
2(log(t))2
Consider the homogeneous DE with FICs.
(t2D2−3tD+4)x = 0
x(1) = (−0.5,0,1)
x′(1) = (−0.5,0,0.5)
The problem is homogeneous and initial values are fuzzy numbers with vertices at 0.
Therefore the solution by solution algorithm
x1(t) = t2 and x2(t) = t2 log(t) are linearly independent solutions of
(t2D2−3tD+4)y = 0. Hence s(t) = (t2, t2 log(t)) and
W =
∣∣∣∣∣∣1 0
2 1
∣∣∣∣∣∣
and y(t) = (y1(t),y2(t)) = ([1−2log(t)]t2, t2 log(t))
The fuzzy solution is
xun = ([1−2log(t)]t2)(−0.5,0,1)+(t2 log(t))(−0.5,0,0.5)
Solutions of nth order Differential Equations with Fuzzy Initial Conditions by using theProperties of Linear Transformations 38
The fuzzy solution x(t) form band in the tx-coordinate space (Fig.3.6).
FIGURE 3.6: The fuzzy solution obtained by the properties of linear transformations
Adding this uncertainty to crisp solution gives the fuzzy solution of the given DE with
FICs
x(t) = xcr(t)+ xun(t) = [1−2log(t)]t2(0.5,1,2)+(t2 log(t))(−0.5,0,0.5)+t2
2(log(t))2
The fuzzy solution x(t) form band in the tx-coordinate space (Fig.3.7).
FIGURE 3.7: The fuzzy solution obtained by the properties of linear transformations,The red line represents the crisp solution
Solution of a higher order differential equation with fuzzy initial condition 39
Example 3.6 Consider the variable coefficient DE with FICs
(t3D3−3t2D2 +6tD−6)x =1t
x(1) = (4,5,6)
x′(1) = (12,13,14.5)
x′′(1) = (9.5,10,11)
Consider corresponding crisp problem.
(t3D3−3t2D2 +6tD−6)x =1t
x(1) = 5
x′(1) = 13
x′′(1) = 10
Associated non-homogeneous problem has the solution
xcr(t) =−13824
t +32824
t2− 6924
t3− 124t
Consider the homogeneous DE with FICs.
(t3D3−3t2D2 +6tD−6)x = 0
x(1) = (−1,0,1)
x′(1) = (−1,0,1.5)
x′′(1) = (−0.5,0,1)
The problem is homogeneous and initial values are fuzzy numbers with vertices at 0.
Therefore, the solution by solution algorithm
x1(t) = t, x2(t) = t2 and x2(t) = t3 are linearly independent solution of
(t3D3−3t2D2 +6tD−6)x = 0. Hence s(t) = (t, t2, t3) and
W =
∣∣∣∣∣∣∣∣∣
1 1 1
1 2 3
0 2 6
∣∣∣∣∣∣∣∣∣
Solutions of nth order Differential Equations with Fuzzy Initial Conditions by using theProperties of Linear Transformations 40
and
y(t) = (y1(t),y2(t),y3(t))
=
(3t−3t2 + t3,−2t +3t2− t3,
t2− t2 +
t3
2
)
The fuzzy solution is
xun(t) = (3t−3t2 + t3)(−1,0,1)+(−2t +3t2− t3)(−1,0,1.5)+(
t2− t2 +
t3
2
)(−0.5,0,1)
The fuzzy solution x(t) form band in the tx-coordinate space (Fig.3.8).
FIGURE 3.8: The fuzzy solution obtained by properties of linear transformation
We add this uncertainty to crisp solution and get the fuzzy solution of the given DE with
FICs
x(t) = xcr(t)+ xun(t)
= (3t−3t2 + t3)(4,5,6)+(−2t +3t2− t3)(12,13,14.5)+(
t2− t2 +
t3
2
)(9.5,10,11)− 1
24t
The fuzzy solution x(t) form band in the tx-coordinate space and the red line represents
the crisp solution (Fig.3.9).
Solution of a higher order differential equation with fuzzy initial condition 41
FIGURE 3.9: The fuzzy solution obtained by properties of linear transformation, Thered line represents the crisp solution
In above example, if we increase t and as we take t → ∞ then the fuzziness in the
solution increases and goes to infinite (Fig.3.10).
FIGURE 3.10: If we increase t fuzziness increases
Consider the mass-spring system with damping coefficient zero
mx′′(t)+ kx(t) = r(t),
where r(t) is an external force and the resultant motion is a forced motion.
Example 3.7 Consider a unit mass sliding on a frictionless table attached to a spring,
with spring constant k = 16. Suppose the mass is lightly tapped by a hammer every T
seconds. Mass tapped periodically with hammer. Suppose that the first tap occurs at
Solutions of nth order Differential Equations with Fuzzy Initial Conditions by using theProperties of Linear Transformations 42
time t = 0 and before that time the mass is at rest. Describe what happens to the
motion of the mass for the tapping period T = 1.
FIGURE 3.11: Mass tapped periodically with a hammer
The non-homogeneous DE with initial conditions
x′′(t)+16 x(t) = sin t,
x(0) = 0
x′(0) = 1
Consider the homogeneous equation with fuzzy parameters.
x′′(t)+16 x(t) = 0,
x(0) = (−1,0,1)
x′(0) = (−0.5,0,1)
The problem is homogeneous and initial values are fuzzy numbers with vertex 0.
x1(t) = cos4t, x2(t) = sin4t are linearly independent solutions for the equation
x′′(t)+16 x(t) = 0. Hence s(t) = (cos4t, sin4t) and
W =
∣∣∣∣∣∣1 0
0 4
∣∣∣∣∣∣
and y(t) = (y1(t),y2(t)) =(
cos4t,14
sin4t)
The fuzzy solution of the given homogeneous problem is
xun(t) = (cos4t)(−1,0,1)+14
sin4t(−0.5,0,1) (3.4)
Result and Discussion 43
The crisp solution of the given non-homogeneous problem is
xcr(t) =730
sin4t +1
15sin t (3.5)
If we add uncertainty of (3.4) to crisp solution i.e.(3.5) then we get the fuzzy solution
of the given DE with FICs
x(t) = xcr(t)+ xun(t)
= cos4t(−1,0,1)+7
30sin4t(0.5,1,2)+
115
sin t
where arithmetic operations are considered to be fuzzy operations.
The fuzzy solution x(t) form band in the tx-coordinate space (Fig.3.12).
FIGURE 3.12: The fuzzy solution obtained by the properties of linear transformation,The red line represents the crisp solution
3.6 Result and Discussion
Here, we solved the homogeneous DE with FICs that have both the lower-bounds and
upper bounds of solution (Fig.3.1) and the non-homogeneous DE with crisp initial
conditions that gives us the crisp solution i.e. denoted by the red line in Fig.3.2. If we
add both the solutions we can find the fuzzy solution of the given DE with FICs as in
Solutions of nth order Differential Equations with Fuzzy Initial Conditions by using theProperties of Linear Transformations 44
Examples 3.1 and 3.2. Similarly, we can observe the solution bounds and the crisp
solutions for Examples 3.3 to 3.6. We also solved one application-level problem
related to physics (Example 3.7) for the study of the motion of mass when the mass is
tapped periodically by a hammer.
3.7 Conclusion
The DEs with FICs are investigated as a set of crisp problems. The method based on the
properties of linear transformations is applied for the second and the third-order DEs
with FICs for clarity purpose, but it is also applicable for the nth order DEs with FICs
[66]. By using the properties of linear transformations, we solved homogeneous and
non-homogeneous constant coefficient as well as Cauchy-Euler’s variable coefficient
DEs with FICs.
CHAPTER 4
Solutions of the nth order Differential
Equations with Fuzzy Initial and
Boundary Conditions by using the
Gauss Elimination Method
4.1 Introduction
Differential equations play an important role in modelling the real world problems.
Today, DEs represent a fundamental mathematical tool for studying systems that
change with time, and are used in most of the areas of science and engineering. Hence,
engineering and science students need to be able to model and solve problems using
DEs; thus allowing them to analyse the behaviour of their underlying dynamics.
Generally, the description of a phenomenon is completed with certainty over all its
parameters, including initial and boundary conditions. So, a phenomenon is usually
modelled by DEs and predetermined initial and boundary conditions. Chen et al. [28]
worked on fuzzy boundary value problems. They demonstrated two point boundary
value problems of undamped uncertain dynamical systems[27]. Allahviranloo et al.
[11] offered the method for solving the nth order fuzzy linear differential equations.
Fuzzy approximate solutions of second order fuzzy linear boundary value problems
were obtained by Xiaobin et al. [78]. A boundary value problem for the second order
fuzzy differential equation was studied by Khastan et al. [50]. Solutions of linear
differential equation with fuzzy boundary values were discussed by Gasilov et al. [35].
45
Solutions of the nth order Differential Equations with Fuzzy Initial and BoundaryConditions by using the Gauss Elimination Method 46
O’Regan et al. [62] proposed initial and boundary value problems for fuzzy differential
equations. Here, we consider the DEs with fuzzy parameters and solve them by using
GEM.
4.2 Problem formulation
Here, we shall convert the DE with FIBCs into the system of linear equations, apply
GEM then use MATLAB programming to solve it. We assume that the approximate
solution to the given DE with FIBCs in the form of a polynomial of power k, where
k = 0,1,2, ....,N. We then solve the nth order linear DE with FIBCs by GEM. We also
compare the approximate and the exact solutions and represent them graphically by
using the Mathematica software.
4.3 Method for solving the nth order differential
equation with FBCs
Consider the second-order DE
y′′+ p(t)y′+q(t)y = g(t), a≤ t ≤ b (4.1)
with fuzzy boundary conditions:
y(a) = α, y(b) = β , (4.2)
where p(t) and q(t) are constants or functions of t. Here, we have to find fuzzy
coefficients for approximating the solution as
yN(t) =N
∑k=0
akφk(t) (4.3)
where φk(t), k = 0,1.....,N, are positive polynomial functions whose all derivatives are
positive. Here, we find the coefficients which are fuzzy in (4.3) by considering errors
Method for solving the nth order differential equation with FBCs 47
to zero as follows:
E = D(y′′+ p(t)y
′+q(t)y, g(t))+D(y(a), α)+D(y(b), β ) (4.4)
Then, substitute (4.3) in (4.4) and the parametric form as under
y′′+ p(t)y
′(t,r)+q(t)y(t,r) = g(t,r)
y(a,r) = α(r)
y(b,r) = β (r)
y′′+ p(t)y′(t,r)+q(t)y(t,r) = g(t,r)
y(a,r) = α(r)
y(b,r) = β (r)
(4.5)
Case:1 p(t)q(t)≥ 0
Suppose that the coefficients p(t),q(t) are non-negative. From (4.5), the equations are
y′′(t,r)+ p(t)y
′(t,r)+q(t)y(t,r) = g(t,r) (4.6)
y′′(t,r)+ p(t)y
′(t,r)+q(t)y(t,r) = g(t,r) (4.7)
y(a,r) = α(r), y(b,r) = β (r), y(a,r) = α(r), y(b,r) = β (r). (4.8)
Equation (4.3) is substituted in (4.6), (4.7) and (4.8) respectively, then
N
∑k=0
ak(r)φ′′k (t)+ p(t)
N
∑k=0
ak(r)φ′k(t)+q(t)
N
∑k=0
ak(r)φk(t) = g(t,r) (4.9)
N
∑k=0
ak(r)φ′′k (t)+ p(t)
N
∑k=0
ak(r)φ′k(t)+q(t)
N
∑k=0
ak(r)φk(t) = g(t,r) (4.10)
and
N
∑k=0
ak(r)φk(a) = α(r),N
∑k=0
ak(r)φk(b) = β (r) (4.11)
N
∑k=0
ak(r)φk(a) = α(r),N
∑k=0
ak(r)φk(b) = β (r) (4.12)
Solutions of the nth order Differential Equations with Fuzzy Initial and BoundaryConditions by using the Gauss Elimination Method 48
By considering
γk = φ′′k (t)+ p(t)φ
′k(t)+q(t)φk(t),k = 0,1, .....,N
σak = φk(a),σbk = φk(b),
the following system is obtained:
N∑
k=0ak(r)γk = g(t,r)
N∑
k=0ak(r)σak = α(r)
N∑
k=0ak(r)σbk = β (r)
N∑
k=0ak(r)γk = g(t,r)
N∑
k=0ak(r)σak = α(r)
N∑
k=0ak(r)σbk = β (r)
(4.13)
(4.13) is a system of linear equations A(t)X(r) = B(r) such that
A =
A1 A2
A2 A1
,
where
A1 =
γ0 γ1 γ2 γ3
σa0 σa1 σa2 σa3
σb0 σb1 σb2 σb3
, A2 =
0 0 0 0
0 0 0 0
0 0 0 0
and
X = (a0,a1,a2,a3,a0,a1,a2,a3)T , B = (g(t,r),α(r),β (r),g(t,r),α(r),β (r))T .
The parameters a0,a1,a2,a3,a0,a1,a2,a3 are obtained by solving (4.13) and taking
t = s,s ∈ [a,b]. These parameters gives us the fuzzy approximate solution
(y(t,r),y(t,r)).
Method for solving the nth order differential equation with FBCs 49
The system of linear equations A(t)X(r) = B(r) is
γ0 γ1 γ2 γ3 0 0 0 0
σa0 σa1 σa2 σa3 0 0 0 0
σb0 σb1 σb2 σb3 0 0 0 0
0 0 0 0 γ0 γ1 γ2 γ3
0 0 0 0 σa0 σa1 σa2 σa3
0 0 0 0 σb0 σb1 σb2 σb3
a0(r)
a1(r)
a2(r)
a3(r)
a0(r)
a1(r)
a2(r)
a3(r)
=
g(t,r)
α(r)
β (r)
g(t,r)
α(r)
β (r)
Suppose that coefficients p(t),q(t) are negative. Then, from (4.5), the equations are
y′′(t,r)− p(t)y
′(t,r)−q(t)y(t,r) = g(t,r) (4.14)
y′′(t,r)− p(t)y
′(t,r)−q(t)y(t,r) = g(t,r) (4.15)
y(a,r) = α(r), y(b,r) = β (r), y(a,r) = α(r), y(b,r) = β (r). (4.16)
Equation (4.3) when substituted (4.14), (4.15) and (4.16) gives, respectively,
N
∑k=0
ak(r)φ′′k (t)− p(t)
N
∑k=0
ak(r)φ′k(t)−q(t)
N
∑k=0
ak(r)φk(t) = g(t,r) (4.17)
N
∑k=0
ak(r)φ′′k (t)− p(t)
N
∑k=0
ak(r)φ′k(t)−q(t)
N
∑k=0
ak(r)φk(t) = g(t,r) (4.18)
and
N
∑k=0
ak(r)φk(a) = α(r),N
∑k=0
ak(r)φk(b) = β (r) (4.19)
N
∑k=0
ak(r)φk(a) = α(r),N
∑k=0
ak(r)φk(b) = β (r) (4.20)
Solutions of the nth order Differential Equations with Fuzzy Initial and BoundaryConditions by using the Gauss Elimination Method 50
By taking
γk = φ′′k (t)
δk =−p(t)φ′k(t)−q(t)φk(t),k = 0,1, .....,N
σak = φk(a),σbk = φk(b)
the corresponding system of linear equations A(t)X(r) = B(r) is
γ0 γ1 γ2 γ3 δ0 δ1 δ2 δ3
σa0 σa1 σa2 σa3 0 0 0 0
σb0 σb1 σb2 σb3 0 0 0 0
δ0 δ1 δ2 δ3 γ0 γ1 γ2 γ3
0 0 0 0 σa0 σa1 σa2 σa3
0 0 0 0 σb0 σb1 σb2 σb3
a0(r)
a1(r)
a2(r)
a3(r)
a0(r)
a1(r)
a2(r)
a3(r)
=
g(t,r)
α(r)
β (r)
g(t,r)
α(r)
β (r)
Case:2 p(t)q(t)< 0
Suppose that the coefficient p(t) is non-negative and q(t) is negative. From (4.5), the
set of equations are
y′′(t,r)+ p(t)y
′(t,r)−q(t)y(t,r) = g(t,r) (4.21)
y′′(t,r)+ p(t)y
′(t,r)−q(t)y(t,r) = g(t,r) (4.22)
y(a,r) = α(r), y(b,r) = β (r), y(a,r) = α(r), y(b,r) = β (r). (4.23)
Equation (4.3) when substituted in (4.21), (4.22) and (4.23) gives, respectively,
N
∑k=0
ak(r)φ′′k (t)+ p(t)
N
∑k=0
ak(r)φ′k(t)−q(t)
N
∑k=0
ak(r)φk(t) = g(t,r) (4.24)
N
∑k=0
ak(r)φ′′k (t)+ p(t)
N
∑k=0
ak(r)φ′k(t)−q(t)
N
∑k=0
ak(r)φk(t) = g(t,r) (4.25)
Method for solving the nth order differential equation with FBCs 51
and
N
∑k=0
ak(r)φk(a) = α(r),N
∑k=0
ak(r)φk(b) = β (r) (4.26)
N
∑k=0
ak(r)φk(a) = α(r),N
∑k=0
ak(r)φk(b) = β (r) (4.27)
By considering
ζk = φ′′k (t)+ p(t)φ
′k(t)
ξk =−q(t)φk(t),k = 0,1, .....,N
σak = φk(a),σbk = φk(b)
the following system is obtained:
N∑
k=0ak(r)ζk−
N∑
k=0ak(r)ξk = g(t,r)
N∑
k=0ak(r)σak = α(r)
N∑
k=0ak(r)σbk = β (r)
N∑
k=0ak(r)ζk−
N∑
k=0ak(r)ξk = g(t,r)
N∑
k=0ak(r)σak = α(r)
N∑
k=0ak(r)σbk = β (r)
(4.28)
(4.28) is a system of linear equations A(t)X(r) = B(r) such that
A =
A1 A2
A2 A1
,
where
A1 =
ζ0 ζ1 ζ2 ζ3
σa0 σa1 σa2 σa3
σb0 σb1 σb2 σb3
, A2 =
ξ0 ξ1 ξ2 ξ3
0 0 0 0
0 0 0 0
Solutions of the nth order Differential Equations with Fuzzy Initial and BoundaryConditions by using the Gauss Elimination Method 52
Similarly, when p(t) is negative and q(t) is non-negative, from (4.5), the set of
equations are:
y′′(t,r)− p(t)y
′(t,r)+q(t)y(t,r) = g(t,r) (4.29)
y′′(t,r)− p(t)y
′(t,r)+q(t)y(t,r) = g(t,r) (4.30)
y(a,r) = α(r), y(b,r) = β (r), y(a,r) = α(r), y(b,r) = β (r). (4.31)
Equation (4.3) when substituted in (4.29), (4.30) and (4.31) gives, respectively,
N
∑k=0
ak(r)φ′′k (t)− p(t)
N
∑k=0
ak(r)φ′k(t)+q(t)
N
∑k=0
ak(r)φk(t) = g(t,r) (4.32)
N
∑k=0
ak(r)φ′′k (t)− p(t)
N
∑k=0
ak(r)φ′k(t)+q(t)
N
∑k=0
ak(r)φk(t) = g(t,r) (4.33)
and
N
∑k=0
ak(r)φk(a) = α(r),N
∑k=0
ak(r)φk(b) = β (r) (4.34)
N
∑k=0
ak(r)φk(a) = α(r),N
∑k=0
ak(r)φk(b) = β (r) (4.35)
By setting
ζk = φ′′k (t)+q(t)φk(t)
ξk =−p(t)φ′k(t),k = 0,1, .....,N
σak = φk(a),σbk = φk(b)
Method for solving the nth order differential equation with FBCs 53
the following system is obtained:
N∑
k=0ak(r)ζk−
N∑
k=0ak(r)ξk = g(t,r)
N∑
k=0ak(r)σak = α(r)
N∑
k=0ak(r)σbk = β (r)
N∑
k=0ak(r)ζk−
N∑
k=0ak(r)ξk = g(t,r)
N∑
k=0ak(r)σak = α(r)
N∑
k=0ak(r)σbk = β (r)
(4.36)
where
X = (a0,a1,a2,a3,a0,a1,a2,a3)T , B = (g(t,r),α(r),β (r),g(t,r),α(r),β (r))T .
The parameters a0,a1,a2,a3,a0,a1,a2,a3 are obtained by solving (4.28) and (4.36) by
taking t = s,s ∈ [a,b]. These parameters provide the fuzzy approximate solution
(y(t,r),y(t,r)).
The system of linear equations A(t)X(r) = B(r) is
ζ0 ζ1 ζ2 ζ3 ξ0 ξ1 ξ2 ξ3
σa0 σa1 σa2 σa3 0 0 0 0
σb0 σb1 σb2 σb3 0 0 0 0
ξ0 ξ1 ξ2 ξ3 ζ0 ζ1 ζ2 ζ3
0 0 0 0 σa0 σa1 σa2 σa3
0 0 0 0 σb0 σb1 σb2 σb3
a0(r)
a1(r)
a2(r)
a3(r)
a0(r)
a1(r)
a2(r)
a3(r)
=
g(t,r)
α(r)
β (r)
g(t,r)
α(r)
β (r)
Thus, we obtained the fuzzy approximate solution of the given second order DE with
FBCs as
y(t,r) = a0(r)φ0(t)+a1(r)φ1(t)+a2(r)φ2(t)+a3(r)φ3(t) (4.37)
y(t,r) = a0(r)φ0(t)+a1(r)φ1(t)+a2(r)φ2(t)+a3(r)φ3(t) (4.38)
Solutions of the nth order Differential Equations with Fuzzy Initial and BoundaryConditions by using the Gauss Elimination Method 54
In general, for the Nth order DE with FIBCs, the system of linear equations
A(t)2N×2(N+1)X(r)2(N+1)×1 = B(r)2N×1, where N = n+1 is:
ζ0 ζ1 · · · ζN ξ0 ξ1 · · · ξN
σa0 σa1 · · · σaN 0 0 · · · 0
σb0 σb1 · · · σbN 0 0 · · · 0
ξ0 ξ1 · · · ξN ζ0 ζ1 · · · ζN
0 0 · · · 0 σa0 σa1 · · · σaN
0 0 · · · 0 σb0 σb1 · · · σbN
a0(r)
a1(r)...
aN(r)
a0(r)
a1(r)...
aN(r)
=
g(t,r)
α1(r)...
αN(r)
g(t,r)
α1(r)...
αN(r)
and
X =(a0,a1, .....,aN ,a0,a1, ....,aN)T , B=(g(t,r),α1(r), ..,αN(r),g(t,r),α1(r), ..,αN(r))T .
The system of linear equations A(t)X(r) = B(r), by taking t = s,s ∈ [a,b]. Here, we
obtained required solution by GEM by the help of MATLAB programming i.e.
X(r) = A−1(s)B(r)
The parameters a0,a1, .....,aN ,a0,a1, ....,aN are obtained by solving the above system.
These parameters provide the fuzzy approximate solution (y(t,r),y(t,r)).
Thus, we obtained the fuzzy approximate solution of the given Nth order DE with
FIBCs as
y(t,r) = a0(r)φ0(t)+a1(r)φ1(t)+ ....+aN(r)φN(t) (4.39)
y(t,r) = a0(r)φ0(t)+a1(r)φ1(t)+ ....+aN(r)φN(t) (4.40)
Solutions of higher order differential equations with fuzzy initial and boundaryconditions 55
4.4 Solutions of higher order differential equations with
fuzzy initial and boundary conditions
4.4.1 Solutions of a second order differential equation with fuzzy
boundary conditions
Example 4.1 Consider the second-order variable coefficient BVP:
(t2 +1)y′′−2ty′+2y = 0, t ∈ [1,2]
y(1) = 0
y(2) = 3
The general solution i.e. exact solution is
y(t) = c1t + c2(t2−1) (4.41)
By applying the boundary conditions, a particular solution is
y(t) = t2−1 (4.42)
Consider the second-order DE with FBCs:
(t2 +1)y′′−2ty
′+2y = 0 = [−0.25+0.25r,0.25−0.25r], t ∈ [1,2]
y(1) = [−0.5+0.5r,0.5−0.5r]
y(2) = [2.1+0.9r,3.5−0.5r]
The lower and upper bounds of the exact solutions are:
Y (t,r) = (−0.5+0.5r)t +13(3.1−0.1r)(t2−1)
Y (t,r) = (0.5−0.5r)t +13(2.5+0.5r)(t2−1)
Solutions of the nth order Differential Equations with Fuzzy Initial and BoundaryConditions by using the Gauss Elimination Method 56
By using MATLAB calculations
TABLE 4.1: The exact values for lower-bounds.
tr 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0
0.0 -0.5000 -0.3330 -0.1453 0.0630 0.2920 0.5417 0.8120 1.1030 1.4147 1.7470 2.10000.1 -0.4500 -0.2787 -0.0868 0.1257 0.3588 0.6125 0.8868 1.1817 1.4972 1.8333 2.19000.2 -0.4000 -0.2244 -0.0283 0.1884 0.4256 0.6833 0.9616 1.2604 1.5797 1.9196 2.28000.3 -0.3500 -0.1701 0.0303 0.2511 0.4924 0.7542 1.0364 1.3391 1.6623 2.0059 2.37000.4 -0.3000 -0.1158 0.0888 0.3138 0.5592 0.8250 1.1112 1.4178 1.7448 2.0922 2.46000.5 -0.2500 -0.0615 0.1473 0.3765 0.6260 0.8958 1.1860 1.4965 1.8273 2.1785 2.55000.6 -0.2000 -0.0072 0.2059 0.4392 0.6928 0.9667 1.2608 1.5752 1.9099 2.2648 2.64000.7 -0.1500 0.0471 0.2644 0.5019 0.7596 1.0375 1.3356 1.6539 1.9924 2.3511 2.73000.8 -0.1000 0.1014 0.3229 0.5646 0.8264 1.1083 1.4104 1.7326 2.0749 2.4374 2.82000.9 -0.0500 0.1557 0.3815 0.6273 0.8932 0.1792 1.4852 1.8113 2.1575 2.5237 2.91001.0 0.0000 0.2100 0.4400 0.6900 0.9600 1.2500 1.5600 1.8900 2.2400 2.6100 3.0000
TABLE 4.2: The exact values for upper-bounds.
tr 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0
0.0 0.5000 0.7250 0.9667 1.2250 1.5000 1.7917 2.1000 2.4250 2.7667 3.1250 3.50000.1 0.4500 0.6735 0.9140 1.1715 1.4460 1.7375 2.0460 2.3715 2.7140 3.0735 3.45000.2 0.4000 0.6220 0.8613 1.1180 1.3920 1.6833 1.9920 2.3180 2.6613 3.0220 3.40000.3 0.3500 0.5705 0.8087 1.0645 1.3380 1.6292 1.9380 2.2645 2.6087 2.9705 3.35000.4 0.3000 0.5190 0.7560 1.0110 1.2840 1.5750 1.8840 2.2110 2.5560 2.9190 3.30000.5 0.2500 0.4675 0.7033 0.9575 1.2300 1.5208 1.8300 2.1575 2.5033 2.8675 3.25000.6 0.2000 0.4160 0.6507 0.9040 1.1760 1.4667 1.7760 2.1040 2.4507 2.8160 3.20000.7 0.1500 0.3645 0.5980 0.8505 1.1220 1.4125 1.7220 2.0505 2.3980 2.7645 3.15000.8 0.1000 0.3130 0.5453 0.7970 1.0680 1.3583 1.6680 1.9970 2.3453 2.7130 3.10000.9 0.0500 0.2615 0.4927 0.7435 1.0140 1.3042 1.6140 1.9435 2.2927 2.6615 3.05001.0 0.0000 0.2100 0.4400 0.6900 0.9600 1.2500 1.5600 1.8900 2.2400 2.6100 3.0000
Plotting graph using Mathematica software
0.0
0.5
1.0
r
1.0
1.5
2.0
t
0
1
2
3
YHt,rL
FIGURE 4.1: The lower-bound and upper-bound for exact solution
The approximate solution of the given DE is obtained as under:
If φk(t) = tk; where k = 0,1,2,3
Solutions of higher order differential equations with fuzzy initial and boundaryconditions 57
y(t,r) = a0(r)+a1(r)t +a2(r)t2 +a3(r)t
3
y(t,r) = a0(r)+a1(r)t +a2(r)t2 +a3(r)t3
Consider the system of equations
(t2 +1)[2a2(r)+6ta3(r)]−2t[a1(r)+2ta2(r)+3t2a3(r)]
+2[a0(r)+a1(r)t +a2(r)t2 +a3(r)t
3] =−0.25+0.25r
a0(r)+a1(r)+a2(r)+a3(r) =−0.5+0.5r
a0(r)+2a1(r)+4a2(r)+8a3(r) = 2.1+0.9r
(t2 +1)[2a2(r)+6ta3(r)]−2t[a1(r)+2ta2(r)+3t2a3(r)]
+2[a0(r)+a1(r)t +a2(r)t2 +a3(r)t3] = 0.25−0.25r
a0(r)+a1(r)+a2(r)+a3(r) = 0.5−0.5r
a0(r)+2a1(r)+4a2(r)+8a3(r) = 3.5−0.5r
The corresponding linear system is
2 2t 2+4t2 6t +8t3 0 −2t −4t2 −6t3
1 1 1 1 0 0 0 0
1 2 4 8 0 0 0 0
0 −2t −4t2 −6t3 2 2t 2+4t2 6t +8t3
0 0 0 0 1 1 1 1
0 0 0 0 1 2 4 8
a0(r)
a1(r)
a2(r)
a3(r)
a0(r)
a1(r)
a2(r)
a3(r)
=
−0.25+0.25r
−0.5+0.5r
2.1+0.9r
0.25−0.25r
0.5−0.5r
3.5−0.5r
By solving above system we can find approximate solution of the given problem.
Taking t = 1.5, the approximate solution is obtained as follows:
The lower bounds of solution are as:
a0(r) = 7.0833−8.0833r
a1(r) = 17.875r−17.875
a2(r) = 12.8917−11.8917r
a3(r) = 2.6r−2.6
Solutions of the nth order Differential Equations with Fuzzy Initial and BoundaryConditions by using the Gauss Elimination Method 58
y(t,r)= (7.0833−8.0833r)+(17.875r−17.875)t+(12.8917−11.8917r)t2+(2.6r−2.6)t3
The upper bounds of solution are as:
a0(r) = −0.5r−0.5
a1(r) = 0
a2(r) = 1
a3(r) = 0
y(t,r) = (−0.5r−0.5)+ t2
By using MATLAB calculations
TABLE 4.3: The approximate values for lower-bounds.
tr 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0
0.0 -0.5000 -0.4409 -0.2955 -0.0795 0.1961 0.5021 0.8364 1.1789 1.5141 1.8263 2.10000.1 -0.4500 -0.3758 -0.2219 -0.0025 0.2684 0.5769 0.9088 1.2501 1.5867 1.9047 2.19000.2 -0.4000 -0.3107 -0.1484 0.0744 0.3453 0.6517 0.9811 1.3212 1.6593 1.9831 2.28000.3 -0.3500 -0.2456 -0.0784 0.1514 0.4221 0.7265 1.0535 1.3923 1.7319 2.0614 2.37000.4 -0.3000 -0.1805 -0.0013 0.2283 0.4940 0.8013 1.1258 1.4634 1.8045 2.1398 2.46000.5 -0.2500 -0.1154 0.0723 0.3053 0.5758 0.8760 1.1982 1.5345 1.8771 2.2182 2.55000.6 -0.2000 -0.0503 0.1458 0.3822 0.6526 0.9508 1.2706 1.6056 1.9497 2.2965 2.64000.7 -0.1500 0.0147 0.2194 0.4592 0.7295 1.0256 1.3429 1.6767 2.0222 2.3749 2.73000.8 -0.1000 0.0798 0.2929 0.5361 0.8063 1.1004 1.4153 1.7478 2.0948 2.4533 2.82000.9 -0.0500 0.1449 0.3665 0.6131 0.8832 0.1752 1.4876 1.8189 2.1674 2.5316 2.91001.0 0.0000 0.2100 0.4400 0.6900 0.9600 1.2500 1.5600 1.8900 2.2400 2.6100 3.0000
TABLE 4.4: The approximate values for upper-bounds.
tr 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0
0.0 0.5000 0.7100 0.9400 1.1900 1.4600 1.7500 2.0600 2.3900 2.7400 3.1100 3.50000.1 0.4500 0.6600 0.8900 1.1400 1.4100 1.7000 2.0100 2.3400 2.6900 3.0600 3.45000.2 0.4000 0.6100 0.8400 1.0900 1.3600 1.6500 1.9600 2.2900 2.6400 3.0100 3.40000.3 0.3500 0.5600 0.7900 1.0400 1.3100 1.6000 1.9100 2.2400 2.5900 2.9600 3.35000.4 0.3000 0.5100 0.7400 0.9900 1.2600 1.5500 1.8600 2.1900 2.5400 2.9100 3.30000.5 0.2500 0.4600 0.6900 0.9400 1.2100 1.5000 1.8100 2.1400 2.4900 2.8600 3.25000.6 0.2000 0.4100 0.6400 0.8900 1.1600 1.4500 1.7600 2.0900 2.4400 2.8100 3.20000.7 0.1500 0.3600 0.5900 0.8400 1.1100 1.4000 1.7100 2.0400 2.3900 2.7600 3.15000.8 0.1000 0.3100 0.5400 0.7900 1.0600 1.3500 1.6600 1.9900 2.3400 2.7100 3.10000.9 0.0500 0.2600 0.4900 0.7400 1.0100 1.3000 1.6100 1.9400 2.2900 2.6600 3.05001.0 0.0000 0.2100 0.4400 0.6900 0.9600 1.2500 1.5600 1.8900 2.2400 2.6100 3.0000
Solutions of higher order differential equations with fuzzy initial and boundaryconditions 59
Plotting graph using Mathematica software
0.0
0.5
1.0
r
1.0
1.5
2.0
t
0
1
2
3
yHt,rL
FIGURE 4.2: The lower-bound and upper-bound for approximate solution
4.4.2 Solution of fourth order differential equation with fuzzy
initial and boundary conditions
4.4.2.1 Example based on the application of differential equation in civil
engineering
Consider a uniform beam as made up of fibres running lengthwise. We have to find its
deflection under given loadings.
M =EIR
Where R is a radius of curvature of elastic curve.
R =[1+(dy
dx)2]
32
d2ydx2
Thus for small deflection R =1
d2ydx2
Bending moment M = EId2ydx2
Shear force(=
dMdx
)= EI
d3ydx3
Intensity of loading(=
d2Mdx2
)= EI
d4ydx4
Solutions of the nth order Differential Equations with Fuzzy Initial and BoundaryConditions by using the Gauss Elimination Method 60
Example 4.2 A beam of length L is embedded at both ends. Find the deflection of the
beam if a constant load w0 is uniformly distributed along its length that is,
w(x) = w0,0 < x < L .
EId4ydx4 = w0,x ∈ [0,L]
y(0) = 0,y′(0) = 0,y(L) = 0,y
′(L) = 0
x
y
Wo
θRθL δR1 R2
L
o
FIGURE 4.3: Elastic or Deflection curve
The general solution i.e exact solution is
y(x) = c1 + c2x+ c3x2 + c4x3 +w0
EIx4 (4.43)
By using the boundary conditions, a particular solution is
y(x) =w0L2
24EIx2− w0L2
12EIx3 +
w0
24EIx4
y(x) =w0
24EIx2(x−L)2
If w0 = 24EI,L = 1 then y(x) = x2(x−1)2
Consider the fourth-order DE with FIBCs:
Here,we take w0 = 24EI,L = 1
d4ydx4 = 24,x ∈ [0,1]
y(0) = y′(0) = y(1) = y
′(1) = [r−1,1− r]
Solutions of higher order differential equations with fuzzy initial and boundaryconditions 61
The lower and upper bounds of the exact solutions are:
Y (x,r) = (r−1)+(r−1)x+[(−6r+6)+
23+ r24
]x2 +
[(4r−4)− 2(23+ r)
24
]x3 +
[23+ r
24
]x4
Y (x,r) = (1− r)+(1− r)x+[(6r−6)+
25− r24
]x2 +
[(−4r+4)− 2(25− r)
24
]x3 +
[25− r
24
]x4
By using MATLAB calculations
TABLE 4.5: The exact values for lower-bounds.
xr 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
0.0 -1.0000 -1.0362 -0.9675 -0.8257 -0.6408 -0.4401 -0.2488 -0.0897 0.0165 0.0518 -0.00000.1 -0.9000 -0.9318 -0.8682 -0.7388 -0.5710 -0.3898 -0.2182 -0.0764 0.0174 0.0474 0.00000.2 -0.8000 -0.8274 -0.7689 -0.6518 -0.5011 -0.3396 -0.1857 -0.0630 0.0183 0.0430 -0.00000.3 -0.7000 -0.7729 -0.6695 -0.5648 -0.4313 -0.2893 -0.1569 -0.0496 0.0193 0.0387 -0.00000.4 -0.6000 -0.6185 -0.5702 -0.4778 -0.3614 -0.2391 -0.1262 -0.0362 0.0202 0.0343 -0.00000.5 -0.5000 -0.5141 -0.4792 -0.3908 -0.2916 -0.1888 -0.0956 -0.0228 0.0211 0.0299 0.00000.6 -0.4000 -0.4096 -0.3716 -0.3038 -0.2218 -0.1385 -0.0650 -0.0094 0.0220 0.0256 0.00000.7 -0.3000 -0.3052 -0.2723 -0.2169 -0.1519 -0.0883 -0.0343 0.0039 0.0229 0.0212 0.00000.8 -0.2000 -0.2008 -0.1730 -0.1299 -0.0821 -0.0380 -0.0037 0.0173 0.0238 0.0168 -0.00000.9 -0.1000 -0.0963 -0.0737 -0.0429 -0.0122 0.0122 0.0270 0.0307 0.0247 0.0125 -0.00001.0 0.0000 0.0081 0.0256 0.0441 0.0576 0.0625 0.0576 0.0441 0.0256 0.0081 0.0000
TABLE 4.6: The exact values for upper-bounds.
xr 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
0.0 1.0000 1.0524 1.0187 0.9139 0.7560 0.5651 0.3640 0.1779 0.0347 -0.0356 0.00000.1 0.9000 0.9480 0.9194 0.8270 0.6862 0.5148 0.3334 0.1646 0.0338 -0.0312 -0.00000.2 0.8000 0.8436 0.8201 0.7400 0.6163 0.4646 0.3027 0.1512 0.0329 -0.0268 0.00000.3 0.7000 0.7391 0.7207 0.6530 0.5465 0.4143 0.2721 0.1378 0.0319 -0.0225 0.00000.4 0.6000 0.6347 0.6214 0.5660 0.4766 0.3641 0.2414 0.1244 0.0310 -0.0181 0.00000.5 0.5000 0.5303 0.5221 0.4790 0.4068 0.3138 0.2108 0.1110 0.0301 -0.0137 0.00000.6 0.4000 0.4258 0.4228 0.3920 0.3370 0.2635 0.1802 0.0976 0.0292 -0.0094 0.00000.7 0.3000 0.3214 0.3235 0.3051 0.2671 0.2133 0.1495 0.0843 0.0283 -0.0050 -0.00000.8 0.2000 0.2170 0.2242 0.2181 0.1973 0.1630 0.1189 0.0709 0.0274 -0.0006 0.00000.9 0.1000 0.1125 0.1249 0.1311 0.1274 0.1128 0.0882 0.0575 0.0265 0.0037 0.00001.0 0.0000 0.0081 0.0256 0.0441 0.0576 0.0625 0.0576 0.0441 0.0256 0.0081 0.0000
Plotting graph using Mathematica software
0.0
0.5
1.0
r
0.0
0.5
1.0
x
-1.0
-0.5
0.0
0.5
1.0
YHx,rL
FIGURE 4.4: The lower-bound and upper-bound for exact solution
Solutions of the nth order Differential Equations with Fuzzy Initial and BoundaryConditions by using the Gauss Elimination Method 62
The approximate solution of the given DE is obtained as under:
If φk(x) = xk where k = 0,1,2,3,4,5
y(x,r) = a0(r)+a1(r)x+a2(r)x2 +a3(r)x
3 +a4(r)x4 +a5(r)x
5
y(x,r) = a0(r)+a1(r)x+a2(r)x2 +a3(r)x3 +a4(r)x4 +a5(r)x5
Consider the system of equations
24a4(r)+120xa5(r) = 23+ r
a0(r) = r−1
a1(r) = r−1
a0(r)+a1(r)+a2(r)+a3(r)+a4(r)+a5(r) = r−1
a1(r)+2a2(r)+3a3(r)+4a4(r)+5a5(r) = r−1
24a4(r)+120xa5(r) = 25− r
a0(r) = 1− r
a1(r) = 1− r
a0(r)+a1(r)+a2(r)+a3(r)+a4(r)+a5(r) = 1− r
a1(r)+2a2(r)+3a3(r)+4a4(r)+5a5(r) = 1− r
The corresponding linear system is
0 0 0 0 24 120x 0 0 0 0 0 0
1 0 0 0 0 0 0 0 0 0 0 0
0 1 0 0 0 0 0 0 0 0 0 0
1 1 1 1 1 1 0 0 0 0 0 0
0 1 2 3 4 5 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 24 120x
0 0 0 0 0 0 1 0 0 0 0 0
0 0 0 0 0 0 0 1 0 0 0 0
0 0 0 0 0 0 1 1 1 1 1 1
0 0 0 0 0 0 0 1 2 3 4 5
a0(r)
a1(r)
a2(r)
a3(r)
a4(r)
a5(r)
a0(r)
a1(r)
a2(r)
a3(r)
a4(r)
a5(r)
=
23+ r
r−1
r−1
r−1
r−1
25− r
1− r
1− r
1− r
1− r
By solving above system we can find approximate solution of the given problem.
Solutions of higher order differential equations with fuzzy initial and boundaryconditions 63
Taking x = 0.5, the approximate solution is obtained as follows:
The lower bounds of solution are as:
a0(r) = a1(r) = r−1
a2(r) = 3.9583−2.9583r
a3(r) = 1.9167r−3.9167
a4(r) = 0.04167r+0.9583
a5(r) = 0
y(x,r)= (r−1)+(r−1)x+(3.9583−2.9583r)x2+(1.9167r−3.9167)x3+(0.04167r+0.9583)x4
The upper bounds of solution are as:
a0(r) = a1(r) = 1− r
a2(r) = 2.9583r−1.9583
a3(r) = −1.9167r−0.0833
a4(r) = 1.04167−0.04167r
a5(r) = 0
y(x,r)= (1−r)+(1−r)x+(2.9583r−1.9583)x2+(−1.9167r−0.0833)x3+(1.04167−0.04167r)x4
By using MATLAB calculations
TABLE 4.7: The approximate values for lower-bounds.
xr 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
0.0 -1.0000 -1.0642 -1.0715 -1.0417 -0.9928 -0.9401 -0.8968 -0.8737 -0.8795 -0.9202 -1.00000.1 -0.9000 -0.9570 -0.9618 -0.9332 -0.8878 -0.8398 -0.8014 -0.7820 -0.7890 -0.8274 -0.90000.2 -0.8000 -0.8498 -0.8521 -0.8246 -0.7827 -0.7396 -0.7059 -0.6902 -0.6985 -0.7346 -0.80000.3 -0.7000 -0.7425 -0.7423 -0.7160 -0.6777 -0.6393 -0.6105 -0.5984 -0.6079 -0.6417 -0.70000.4 -0.6000 -0.6353 -0.6326 -0.6074 -0.5726 -0.5391 -0.5150 -0.5066 -0.5174 -0.5489 -0.60000.5 -0.5000 -0.5281 -0.5229 -0.4988 -0.4676 -0.4388 -0.4196 -0.4148 -0.4269 -0.4561 -0.50000.6 -0.4000 -0.4208 -0.4132 -0.3902 -0.3626 -0.3385 -0.3242 -0.3230 -0.3364 -0.3632 -0.40000.7 -0.3000 -0.3136 -0.3035 -0.2817 -0.2575 -0.2383 -0.2287 -0.2313 -0.2459 -0.2704 -0.30000.8 -0.2000 -0.2064 -0.1938 -0.1731 -0.1525 -0.1380 -0.1333 -0.1395 -0.1554 -0.1776 -0.20000.9 -0.1000 -0.0991 -0.0841 -0.0645 -0.0475 -0.0378 -0.0378 -0.0477 -0.0649 -0.0847 -0.10001.0 0.0000 0.0081 0.0256 0.0441 0.0576 0.0625 0.0576 0.0441 0.0256 0.0081 0.0000
Solutions of the nth order Differential Equations with Fuzzy Initial and BoundaryConditions by using the Gauss Elimination Method 64
TABLE 4.8: The approximate values for upper-bounds.
xr 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
0.0 1.0000 1.0804 1.1227 1.1299 1.1080 1.0651 1.0120 0.9619 0.9307 0.9364 1.00000.1 0.9000 0.9732 0.0130 0.0124 1.0030 0.9648 0.9166 0.8702 0.8402 0.8435 0.90000.2 0.8000 0.8660 0.9033 0.9128 0.8979 0.8646 0.8211 0.7784 0.7497 0.7508 0.80000.3 0.7000 0.7587 0.7935 0.8042 0.7929 0.7643 0.7257 0.6866 0.6591 0.6579 0.70000.4 0.6000 0.6515 0.6838 0.6956 0.6878 0.6641 0.6302 0.5948 0.5686 0.5651 0.60000.5 0.5000 0.5443 0.5741 0.5870 0.5828 0.5638 0.5348 0.5030 0.4781 0.4723 0.50000.6 0.4000 0.4370 0.4644 0.4784 0.4778 0.4635 0.4394 0.4112 0.3876 0.3794 0.40000.7 0.3000 0.3298 0.3547 0.3727 0.3727 0.3633 0.3439 0.3195 0.2971 0.2866 0.30000.8 0.2000 0.2226 0.2450 0.2677 0.2677 0.2630 0.2485 0.2277 0.2066 0.1938 0.20000.9 0.1000 0.1153 0.1353 0.1626 0.1626 0.1628 0.1530 0.1359 0.1161 0.1009 0.10001.0 0.0000 0.0081 0.0256 0.0441 0.0576 0.0625 0.0576 0.0441 0.0256 0.0081 0.0000
Plotting graph using Mathematica software
0.0
0.5
1.0
r
0.0
0.5
1.0
x
-1.0
-0.5
0.0
0.5
1.0
yHx,rL
FIGURE 4.5: The lower-bound and upper-bound for approximate solution
4.5 Result and Discussion
Tables 4.1, 4.2 gives us lower and upper bounds for the exact solution and Tables 4.3,
4.4 gives lower and upper bounds for an approximate solution for Example 4.1 i.e. the
DE with FBCs, respectively. Similarly, the Tables 4.5, 4.6 gives us lower and upper
bounds for the exact solution and the Tables 4.7, 4.8 gives lower and upper bounds for
an approximate solution for Example 4.2 i.e. the non-homogeneous DE which occurred
in civil engineering in that we can estimate the lower and upper bounds of the defection
of the beam under constant load, respectively. In all the above Tables we can observe
that, in between some changes in parameters but for r = 1 endpoints are coincide with
Conclusion 65
given exact conditions of the DEs. Figs. 4.1, 4.2, 4.3, 4.4 shows graphical representation
the lower and upper bounds of the exact and an approximate solution for Examples 4.1,
4.2 respectively. Here, the Figs. 4.1, 4.2 nearly consistent with each other and the Figs.
4.3, 4.4 nearly consistent with each other for Examples 4.1, 4.2 respectively.
4.6 Conclusion
The DEs with FIBCs are investigated as a solution of the matrix. Here the Gauss
elimination method is applied to solve the second-order DE with FBCs and the
fourth-order DE with FIBCs for simplicity purpose but it is also applicable to solve the
nth order DEs with FIBCs. If we assume that, the approximate solution of the problem
in the form of a polynomial of one power more than the order of given DE with FIBCs
then the obtained solution of a given problem is nearer to the exact solution. Here, we
discussed constant-coefficient, variable coefficient, homogeneous and
non-homogeneous higher-order DEs with FIBCs. Approximate and exact solutions of
given examples are nearly consistent with each-other. So, our method is relevant for
solving nth order DEs with FIBCs.
CHAPTER 5
Solutions of Higher Order Differential
Equations with Fuzzy Boundary
Conditions by Finite Difference
Method
5.1 Introduction
Most physical phenomena are modelled by the system of ordinary or partial DEs.
Differential equations with FBCs play an important role in modelling problems in
Science and Engineering because of the way they model dynamical systems under
fuzziness. Opanuga et al. [61] used the finite difference method to solve crisp
boundary value problems. Numerical solutions of FDEs by the predictor-corrector
method were discussed by Bede [16]. An improved predictor-corrector method for
solving fuzzy initial value problems was discussed by Allahviranloo et al. [7]. Ma et
al. [56] used classical Euler’s method to solve FDEs. Friedman et al. [33] proposed a
numerical method for solving fuzzy differential and integral equations. The Euler’s
approximation method was used to solve the FDEs under GHD by Nieto et al. [60].
The Runge-Kutta method was applied by Palligkinis et al. [63] for solving FDEs.
Armand et al. [13] solved two-point FBVP using the variational iteration method.
In general, to find the exact solution of BVPs is difficult, so we have to rely on
numerical methods. Here, we are going to convert DEs with FBCs to the system of
linear equations by using difference equations, and then solve that system of linear
66
Problem formulation 67
equations by FDM and using MATLAB programming. We also solve Airy’s
non-homogeneous DE with FBCs.
5.2 Problem formulation
In mathematics, in the field of DEs, a BVP is a DE together with a set of additional
constrains, called the boundary conditions. The solution to a BVP is a solution to the
DE which also satisfies the boundary conditions.
On the other hand, we are going to solve DEs with FBCs by using finite
difference method. In that, the DE with FBCs on a particular interval is divided into
small intervals and solved using MATLAB. So, in this way, we can apply MATLAB to
approximate the solution of a particular DE with FBCs and represent it graphically to
determine the lower and upper bounds of the solution.
Theorem: 5.2.1 Lipschitz conditions
Suppose the function f in the DE with FBCs
y′′= f (x,y,y
′), a≤ x≤ b, y(a) = α, y(b) = β
is continuous on the set
D = {(x,y,y′)/a≤ x≤ b,−∞ < y < ∞,−∞ < y′< ∞}
and that the partial derivatives fy and f′y are also continuous in D. If
(1) fy(x,y,y′)> 0, ∀ (x,y,y′) ∈ D
(2) a constant M exists with | fy′ (x,y,y′)| ≤M, ∀ (x,y,y′) ∈ D
then the DE with FBCs has a unique solution in terms of parametric forms of fuzzy
numbers.
The finite difference approximation to the various derivatives are as under:
If y(x) and its derivatives are single valued continuous functions of x then by Taylor’s
expansion, we get
Solutions of Higher Order Differential Equations with Fuzzy Boundary Conditions byFinite Difference Method 68
y(x+h) = y(x)+hy′(x)+
h2
2y′′(x)+
h3
6y′′′(x)+
h4
24yiv(x)(ζ+), (5.1)
for some ζ+ ∈ (x,x+h), and
y(x−h) = y(x)−hy′(x)+
h2
2y′′(x)− h3
6y′′′(x)+
h4
24yiv(x)(ζ−), (5.2)
for some ζ− ∈ (x−h,x).
Equation (5.1) gives
y′(x) =
1h[y(x+h)− y(x)]− h
2y′′(x)− ......
i.e.
y′(x) =
1h[y(x+h)− y(x)]+O(h)
which is the forward difference approximation of y′(x) with an error of the order h.
Similarly, (5.2) gives
y′(x) =
1h[y(x)− y(x−h)]+
h2
y′′(x)+ ......
i.e.
y′(x) =
1h[y(x)− y(x−h)]+O(h)
which is the backward difference approximation of y′(x) with an error of the order h.
Subtracting (5.2) from (5.1), we get
y′(x) =
12h
[y(x+h)− y(x−h)]− h2
6y′′′(η),
for some η ∈ (x−h,x+h)
⇒ y′(x) =
12h
[y(x+h)− y(x−h)]+O(h2)
which is the central-difference approximation of y′(x) with an error of the order h2.
Problem formulation 69
The central difference approximation of y′(x) is better than the forward or backward
difference approximation.
Adding (5.1) and (5.2)
y′′(x) =
1h2 [y(x+h)−2y(x)+ y(x−h)]− h2
24[yiv(x)(ζ+)+ yiv(x)(ζ−)]
The Intermediate Value Theorem can be used to simplify this even further.
y′′(x) =
1h2 [y(x+h)−2y(x)+ y(x−h)]− h2
12yiv(x)(ζ )
for some ζ ∈ (x−h,x+h)
⇒ y′′(x) =
1h2 [y(x+h)−2y(x)+ y(x−h)]+O(h2)
which is the central difference approximation of y′′(x) with an error of the order h2 [61].
Similarly we can derive central difference approximations to higher order derivatives.
Thus, the working expressions for the central difference approximations to the first two
derivatives of yi are:
y′i =
12h
(yi+1− yi−1) (5.3)
y′′i =
1h2 (yi+1−2yi + yi−1) (5.4)
Similarly, we can find central difference approximations to higher order derivatives.
y′′′i =
12h3 (yi+2−2yi+1 +2yi−1− yi−2) (5.5)
yivi =
1h4 (yi+2−4yi+1 +6yi−4yi−1 + yi−2) (5.6)
Solutions of Higher Order Differential Equations with Fuzzy Boundary Conditions byFinite Difference Method 70
5.3 Method for solving higher order differential
equation with FBCs
Consider the second order differential equation
y′′(x)+ p(x)y′(x)+q(x)y(x) = r(x), x ∈ [a,b] (5.7)
together with the FBCs
y(a) = α, y(b) = β (5.8)
where p(x),q(x) and r(x) are constants or functions of x on [a,b].
Here, we divide the interval I = [a,b] into a chosen n numbers of subintervals of equal
width. Thus, the step size h of each of the n subintervals is given by h =b−a
n.
yi denotes the value of the function at ith node of the computational grid.
y0 = α, yn = β .
At any mesh point or node point x = xi the finite difference representation of the
differential equation with fuzzy boundaries can be written for as (i = 1,2, ...,n−1):
yi+1−2yi + yi−1
h2 + p(xi)yi+1− yi−1
2h+q(xi)yi = r(xi)+
h2
12[2p(xi)y
′′′(ηi)+ yiv(ζi)].
(5.9)
A Finite difference method with truncation error of the order O(h2) results into the
equation(5.10) after ignoring truncation error [61].
2(yi+1−2yi + yi−1)+hp(xi)(yi+1− yi−1)+2h2q(xi)yi = 2h2r(xi). (5.10)
The boundary conditions provides the solution at the two ends of the grid
i.e. at y0 = α and yn = β .
Method for solving higher order differential equation with FBCs 71
The system of linear equations A(N+1)×(N+1)X(N+1)×1 = B(N+1)×1 is obtained.
A =
1 0 0 · · · · · · 0
(2h2q1−4) (2+hp1) 0 · · · · · · 0
(2−hp2) (2h2q2−4) (2+hp2) 0 · · · 0
0 (2−hp3) (2h2q3−4) (2+hp3) · · · 0...
... . . . . . . . . . ...
0 · · · 0 (2−hpn−2)(2h2qn−2−4) (2+hpn−2)
0 · · · · · · 0 (2−hpn−1) (2h2qn−1−4)
0 0 · · · · · · 0 1
X =
y0
y1
y2
y3...
yn−2
yn−1
yn
B =
α
2h2r1− (2−hp1)α
2h2r2
2h2r3...
2h2rn−2
2h2rn−1− (2+hpn−1)β
β
The above system is a non-singular tridiagonal system and it has a unique solution in
form of fuzzy parameters.
Solutions of Higher Order Differential Equations with Fuzzy Boundary Conditions byFinite Difference Method 72
5.4 Solution of the higher order differential equation
with fuzzy boundary conditions
5.4.1 Solution of the second order differential equation with fuzzy
boundary conditions
Example 5.1 Consider the DE with crisp boundary conditions
y′′= x+ y, x ∈ [0,1]
y(0) = 0
y(1) = 0
Exact solution is
Y (x) =sinhxsinh1
− x. (5.11)
Consider the DE with FBCs
y′′= x+ y, x ∈ [0,1]
y(0) = y(1) = [−0.5+0.5r,0.5−0.5r]
Put xn = x0 +nh, x0 = 0, h = 0.1 and let yn to be calculated as yn = y(xn)
Here, we divide the interval [0,1] into 10 sub-intervals.
By using the central difference approximation:
yi+1−2yi + yi−1
h2 = xi + yi, i = 1,2, ....,10,
for i = 1:
100y0−201y1 +100y2 = 0.1
for i = 2:
100y1−201y2 +100y3 = 0.2
Solution of the higher order differential equation with fuzzy boundary conditions 73
for i = 3:
100y2−201y3 +100y4 = 0.3
and so on
for i = 9:
100y8−201y9 +100y10 = 0.9
We set y0 = y10 = −0.5+ 0.5r,r ∈ [0,1], to obtain the system of linear equations with
unknowns y0,y1,y2, ....,y10 as:
1 0 0 0 0 0 0 0 0 0 0
0−201 100 0 0 0 0 0 0 0 0
0 100 −201 100 0 0 0 0 0 0 0
0 0 100 −201 100 0 0 0 0 0 0
0 0 0 100 −201 100 0 0 0 0 0
0 0 0 0 100 −201 100 0 0 0 0
0 0 0 0 0 100 −201 100 0 0 0
0 0 0 0 0 0 100 −201 100 0 0
0 0 0 0 0 0 0 100 −201 100 0
0 0 0 0 0 0 0 0 100 −201 0
0 0 0 0 0 0 0 0 0 0 1
y0
y1
y2
y3
y4
y5
y6
y7
y8
y9
y10
=
0.5r−0.5
0.6−0.5r
0.2
0.3
0.4
0.5
0.6
0.7
0.8
1.4−0.5r
0.5r−0.5
Solutions of Higher Order Differential Equations with Fuzzy Boundary Conditions byFinite Difference Method 74
The lower-bounds for the solution are:
y0(0) = 0.5r−0.5
y1(0.1) = 0.00479373r−0.019549
y2(0.2) = 0.00463541r−0.0332936
y3(0.3) = 0.00452343r−0.0453711
y4(0.4) = 0.00445669r−0.0549022
y5(0.5) = 0.00443452r−0.0609824
y6(0.6) = 0.00445669r−0.0626725
y7(0.7) = 0.00452343r−0.0589892
y8(0.8) = 0.00463541r−0.0488959
y9(0.9) = 0.00479373r−0.0312915
y10(1) = 0.5r−0.5
Next, by setting y0 = y10 = 0.5− 0.5r,r ∈ [0,1], we get the system of linear equations
with unknowns y0,y1,y2, ....,y10 as:
1 0 0 0 0 0 0 0 0 0 0
0−201 100 0 0 0 0 0 0 0 0
0 100 −201 100 0 0 0 0 0 0 0
0 0 100 −201 100 0 0 0 0 0 0
0 0 0 100 −201 100 0 0 0 0 0
0 0 0 0 100 −201 100 0 0 0 0
0 0 0 0 0 100 −201 100 0 0 0
0 0 0 0 0 0 100 −201 100 0 0
0 0 0 0 0 0 0 100 −201 100 0
0 0 0 0 0 0 0 0 100 −201 0
0 0 0 0 0 0 0 0 0 0 1
y0
y1
y2
y3
y4
y5
y6
y7
y8
y9
y10
=
0.5−0.5r
−0.4+0.5r
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.4+0.5r
0.5−0.5r
Solution of the higher order differential equation with fuzzy boundary conditions 75
which gives the upper-bounds for the solution:
y0(0) = 0.5−0.5r
y1(0.1) = −0.00479373r−0.00996158
y2(0.2) = −0.00463541r−0.0240228
y3(0.3) = −0.00452343r−0.0363242
y4(0.4) = −0.00445669r−0.0459889
y5(0.5) = −0.00443452r−0.0521134
y6(0.6) = −0.00445669r−0.0537591
y7(0.7) = −0.00452343r−0.0499424
y8(0.8) = −0.00463541r−0.0396251
y9(0.9) = −0.00479373r−0.021704
y10(1) = 0.5−0.5r
By using MATLAB, the calculations obtained are as in Tables 5.1 and 5.2.
TABLE 5.1: Approximate values for the lower-bounds.
xr 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
0.0 -0.5000 -0.0195 -0.0333 -0.0454 -0.0549 -0.0610 -0.0627 -0.0590 -0.0489 -0.0313 -0.50000.1 -0.4500 -0.0191 -0.0328 -0.0449 -0.0545 -0.0605 -0.0622 -0.0585 -0.0484 -0.0308 -0.45000.2 -0.4000 -0.0186 -0.0324 -0.0445 -0.0540 -0.0601 -0.0618 -0.0581 -0.0480 -0.0303 -0.40000.3 -0.3500 -0.0181 -0.0319 -0.0440 -0.0536 -0.0597 -0.0613 -0.0576 -0.0475 -0.0299 -0.35000.4 -0.3000 -0.0176 -0.0314 -0.0436 -0.0531 -0.0592 -0.0609 -0.0572 -0.0470 -0.0294 -0.30000.5 -0.2500 -0.0172 -0.0310 -0.0431 -0.0527 -0.0588 -0.0604 -0.0567 -0.0466 -0.0289 -0.25000.6 -0.2000 -0.0167 -0.0305 -0.0427 -0.0522 -0.0583 -0.0600 -0.0563 -0.0461 -0.0284 -0.20000.7 -0.1500 -0.0162 -0.0300 -0.0422 -0.0518 -0.0579 -0.0596 -0.0558 -0.0457 -0.0279 -0.15000.8 -0.1000 -0.0157 -0.0296 -0.0418 -0.0513 -0.0574 -0.0591 -0.0554 -0.0452 -0.0275 -0.10000.9 -0.0500 -0.0152 -0.0291 -0.0413 -0.0509 -0.0570 -0.0587 -0.0549 -0.0447 -0.0270 -0.05001.0 -0.0000 -0.0148 -0.0287 -0.0408 -0.0504 -0.0562 -0.0582 -0.0545 -0.0443 -0.0265 0.0000
Solutions of Higher Order Differential Equations with Fuzzy Boundary Conditions byFinite Difference Method 76
TABLE 5.2: Approximate values for the upper-bounds.
xr 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
0.0 0.5000 -0.0100 -0.0240 -0.0363 -0.0460 -0.0521 -0.0538 -0.0499 -0.0396 -0.0217 0.50000.1 0.4500 -0.0104 -0.0245 -0.0368 -0.0464 -0.0526 -0.0542 -0.0504 -0.0401 -0.0222 0.45000.2 0.4000 -0.0109 -0.0249 -0.0372 -0.0469 -0.0530 -0.0547 -0.0508 -0.0406 -0.0227 0.40000.3 0.3500 -0.0114 -0.0254 -0.0377 -0.0473 -0.0534 -0.0551 -0.0513 -0.0410 -0.0231 0.35000.4 0.3000 -0.0119 -0.0259 -0.0381 -0.0478 -0.0539 -0.0555 -0.0518 -0.0415 -0.0236 0.30000.5 0.2500 -0.0124 -0.0263 -0.0386 -0.0482 -0.0543 -0.0560 -0.0522 -0.0419 -0.0241 0.25000.6 0.2000 -0.0128 -0.0268 -0.0390 -0.0487 -0.0548 -0.0564 -0.0527 -0.0424 -0.0246 0.20000.7 0.1500 -0.0133 -0.0273 -0.0395 -0.0491 -0.0552 -0.0569 -0.0531 -0.0429 -0.0251 0.15000.8 0.1000 -0.0138 -0.0277 -0.0399 -0.0496 -0.0557 -0.0573 -0.0536 -0.0433 -0.0255 0.10000.9 0.0500 -0.0143 -0.0282 -0.0404 -0.0500 -0.0561 -0.0578 -0.0540 -0.0438 -0.0260 0.05001.0 0.0000 -0.0148 -0.0287 -0.0408 -0.0504 -0.0565 -0.0582 -0.0545 -0.0443 -0.0265 0.0000
These bounds are shown graphically as in Fig.5.1.
0-0.5
0.11
0
0.20.9
Y v
alue
s
0.5
0.8 0.30.7 0.40.6 0.5
r values0.5
X values0.60.4 0.70.3 0.80.2 0.90.1 01
FIGURE 5.1: The lower-bounds and the upper-bounds for an approximate solution
Solution of the higher order differential equation with fuzzy boundary conditions 77
5.4.2 Airy’s functions
In physical sciences, the Airy’s function (or Airy’s function of the first kind) Ai(x) is a
special function named after the British astronomer George Biddell Airy (1801–92).
They commonly appear in physics, especially in optics, quantum mechanics,
electromagnetic, and radioactive transfer. The Airy’s stress function is employed in
solid mechanics. The function Ai(x) and the related function Bi(x), are linearly
independent solutions to the differential equation
d2ydx2 − xy = 0 (5.12)
known as Airy’s equation or Stokes equation. This is the simplest second-order linear
differential equation with a turning point (a point where the character of the solutions
changes from oscillatory to exponential). Here, we are going to solve Airy’s
non-homogeneous differential equation with FBCs.
5.4.2.1 Solution of Airy’s non-homogeneous differential equation with fuzzy
boundary conditions
Example 5.2 Consider Airy’s non-homogeneous DE with crisp boundary conditions
y′′−(
1− x5
)y = x, x ∈ [1,3]
y(1) = 2
y(3) =−1
The exact solution is difficult to find and we can use Mathematica/MATLAB to find
the solution, however, it is too long to write and interpret. So, we will first apply FDM
then the obtained system is solved by the help of MATLAB.
Solutions of Higher Order Differential Equations with Fuzzy Boundary Conditions byFinite Difference Method 78
Consider Airy’s non-homogeneous DE with FBCs
y′′−(
1− x5
)y = x, x ∈ [1,3]
y(1) = [1.9+0.1r,2.01−0.01r]
y(3) = [−0.5−0.5r,−1.01+0.01r]
Let xn = x0 +nh, x0 = 1, h = 0.2, and let yn be calculated by yn = y(xn)
Here, we divide the interval [1,3] into 10 sub-intervals.
By using the central difference approximation:
yi+1−2yi + yi−1
h2 −(
1− xi
5
)yi = xi, i = 1,2, ...,10.
This gives,
yi−1−[
2+(
1− xi
5
)h2]
yi + yi+1 = xih2, i = 1,2, ...,10.
For i = 1:
y0−[
2+(
1− 1.25
)(0.2)2
]y1 + y2 = 1.2(0.2)2
for i = 2:
y1−[
2+(
1− 1.45
)(0.2)2
]y2 + y3 = 1.4(0.2)2
for i = 3:
y2−[
2+(
1− 1.65
)(0.2)2
]y3 + y4 = 1.6(0.2)2
and so on
for i = 9:
y8−[
2+(
1− 2.85
)(0.2)2
]y9 + y10 = 2.8(0.2)2
Solution of the higher order differential equation with fuzzy boundary conditions 79
We set y0 = 1.9 + 0.1r, y10 = −0.5− 0.5r, r ∈ [0,1], to obtain the system of linear
equations with unknowns y0,y1,y2, ....y10 as:
1 0 0 0 0 0 0 0 0 0 0
0−2.0304 1 0 0 0 0 0 0 0 0
0 1 −2.0288 1 0 0 0 0 0 0 0
0 0 1 −2.0272 1 0 0 0 0 0 0
0 0 0 1 −2.0256 1 0 0 0 0 0
0 0 0 0 1 −2.0240 1 0 0 0 0
0 0 0 0 0 1 −2.0224 1 0 0 0
0 0 0 0 0 0 1 −2.0208 1 0 0
0 0 0 0 0 0 0 1 −2.0192 1 0
0 0 0 0 0 0 0 0 1 −2.0176 0
0 0 0 0 0 0 0 0 0 0 1
y0
y1
y2
y3
y4
y5
y6
y7
y8
y9
y10
=
1.9+0.1r
−1.852−0.1r
0.056
0.064
0.072
0.08
0.088
0.096
0.1040
0.612+0.5r
−0.5−0.5r
Solutions of Higher Order Differential Equations with Fuzzy Boundary Conditions byFinite Difference Method 80
The lower-bounds for the solution are:
y0(1) = 1.9+0.1r
y1(1.2) = 0.048771r+1.30194
y2(1.4) = 0.791465−0.00097538r
y3(1.6) = 0.359782−0.0507498r
y4(1.8) = 0.00188402−0.101905r
y5(2.0) = −0.213168r−0.49663
y6(2.2) = −0.275443r−0.624419
y7(2.4) = −0.343446r−0.677196
y8(2.6) = 0.359782−0.0507498r
y9(2.8) = −0.418044r−0.638975
y10(3) = −0.5r−0.5
By setting y0 = 2.01− 0.01r, y10 = −1.01+ 0.01r, r ∈ [0,1], we obtained the system
of linear equations with unknowns y0,y1,y2, ...,y10 as:
1 0 0 0 0 0 0 0 0 0 0
0−2.0304 1 0 0 0 0 0 0 0 0
0 1 −2.0288 1 0 0 0 0 0 0 0
0 0 1 −2.0272 1 0 0 0 0 0 0
0 0 0 1 −2.0256 1 0 0 0 0 0
0 0 0 0 1 −2.0240 1 0 0 0 0
0 0 0 0 0 1 −2.0224 1 0 0 0
0 0 0 0 0 0 1 −2.0208 1 0 0
0 0 0 0 0 0 0 1 −2.0192 1 0
0 0 0 0 0 0 0 0 1 −2.0176 0
0 0 0 0 0 0 0 0 0 0 1
y0
y1
y2
y3
y4
y5
y6
y7
y8
y9
y10
Solution of the higher order differential equation with fuzzy boundary conditions 81
=
2.01−0.01r
−1.962+0.01r
0.056
0.064
0.072
0.08
0.088
0.096
0.1040
1.122−0.01r
−1.01+0.01r
The upper-bounds for the solution are:
y0(1) = 2.01−0.01r
y1(1.2) = 1.358362−0.00764766r
y2(1.4) = 0.796018−0.00552781r
y3(1.6) = 0.312599−0.00356715r
y4(1.8) = −0.00170353r−0.0983172
y5(2.0) = 0.000116485r−0.43975
y6(2.2) = 0.0019393r−0.711737
y7(2.4) = 0.00380555r−0.903667
y8(2.6) = 0.00575095r−1.02639
y9(2.8) = 0.00780678r−1.06483
y10(3) = 0.01r−1.01
Solutions of Higher Order Differential Equations with Fuzzy Boundary Conditions byFinite Difference Method 82
Calculations obtained using MATLAB are shown in Tables 5.3 and 5.4.
TABLE 5.3: Approximate values for the lower-bounds.
xr 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.8 3.0
0.0 1.9000 1.3019 0.7915 0.3598 0.0019 -0.2840 -0.4966 -0.6244 -0.6772 -0.6390 -0.50000.1 1.9100 1.3068 0.7914 0.3547 -0.0083 -0.2995 -0.5179 -0.6520 -0.7115 -0.6808 -0.55000.2 1.9200 1.3117 0.7913 0.3496 -0.0185 -0.3151 -0.5393 -0.6795 -0.7459 -0.7226 -0.60000.3 1.9300 1.3166 0.7912 0.3446 -0.0287 -0.3307 -0.5606 -0.7071 -0.7802 -0.7644 -0.65000.4 1.9400 1.3215 0.7911 0.3395 -0.0389 -0.3462 -0.5819 -0.7346 -0.8146 -0.8062 -0.70000.5 1.9500 1.3263 0.7910 0.3344 -0.0491 -0.3618 -0.6032 -0.7621 -0.8489 -0.8480 -0.75000.6 1.9600 1.3312 0.7909 0.3243 -0.0694 -0.3774 -0.6245 -0.7897 -0.8833 -0.8898 -0.80000.7 1.9700 1.3361 0.7908 0.3293 -0.0518 -0.3929 -0.6458 -0.8172 -0.9176 -0.9316 -0.85000.8 1.9800 1.3410 0.7907 0.3192 -0.0796 -0.4085 -0.6672 -0.8448 -0.9520 -0.9734 -0.90000.9 1.9900 1.3458 0.7906 0.3141 -0.0898 -0.4241 -0.6885 -0.8723 -0.9863 -1.0152 -0.95001.0 2.0000 1.3507 0.7905 0.3090 -0.1000 -0.4396 -0.7098 -0.8999 -1.0206 -1.0570 -1.0000
TABLE 5.4: Approximate values for the upper-bounds.
xr 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.8 3.0
0.0 2.0100 1.3584 0.7960 0.3126 -0.0983 -04398 -0.7117 -0.9037 -1.0264 -1.0648 -1.01000.1 2.0090 1.3576 0.7955 0.3122 -0.0985 -04397 -0.7115 -0.9033 -1.0264 -1.0640 -1.00900.2 2.0080 1.3568 0.7949 0.3119 -0.0987 -04397 -0.7113 -0.9029 -1.0252 -1.0633 -1.00800.3 2.0070 1.3561 0.7944 0.3115 -0.0988 -04397 -0.7112 -0.9025 -1.0247 -1.0625 -1.00700.4 2.0060 1.3553 0.7938 0.3112 -0.0990 -04397 -0.7110 -0.0921 -1.0241 -1.0617 -1.00600.5 2.0050 1.3545 0.7933 0.3108 -0.0992 -04397 -0.7108 -0.9018 -1.0235 -1.0609 -1.00500.6 2.0040 1.3538 0.7927 0.3105 -0.0993 -04397 -0.7106 -0.9014 -1.0229 -1.0601 -1.00400.7 2.0030 1.3530 0.7921 0.3101 -0.0995 -04397 -0.7104 -0.9010 -1.0224 -1.0594 -1.00300.8 2.0020 1.3522 0.7916 0.3097 -0.0997 -04397 -0.7102 -0.9006 -1.0218 -1.0586 -1.00200.9 2.0010 1.3515 0.7910 0.3094 -0.0999 -04396 -0.7100 -0.9002 -1.0212 -1.0578 -1.00101.0 2.0000 1.3507 0.7905 0.3090 -1.0000 -04396 -0.7098 -0.8999 -1.0206 -1.0570 -1.0000
Solution of the higher order differential equation with fuzzy boundary conditions 83
These bounds are shown graphically as in Fig.5.2.
-23
-1
2.8
0
2.6
Y v
alue
s 1
2.4 00.1
2
2.2 0.2
X values
2
3
0.30.41.8
r values
0.51.6 0.61.4 0.70.81.2 0.91 1
FIGURE 5.2: The lower-bounds and the upper-bounds for an approximate solution
Solutions of Higher Order Differential Equations with Fuzzy Boundary Conditions byFinite Difference Method 84
5.5 Result and Discussion
Tables 5.1 and 5.2 shows the values of lower and upper bounds for an approximate
solution of Example 5.1 respectively, Fig. 5.1 shows the graphical representation of
these bounds. Table 5.3 and 5.4 shows the values of lower and upper bounds for an
approximate solution of Example 5.2 respectively, Fig. 5.2 shows the graphical
representation of these bounds of Example 5.2 in which Airy’s function is involved. As
we can observe from the above tables, as r increases from 0 to 1 the fuzziness of the
solution decreases, that is, for r = 1, the lower and upper bounds are the same, and that
gives us the required solution.
5.6 Conclusion
Here, we have considered the DEs with FBCs in which after using the difference
equations, we obtain a system which has a mix of fuzzy and crisp parameters on
right-hand side. Upon solving the system using MATLAB, we found all y parameters
to be fuzzy. We also solved higher-order DEs with fuzzy boundaries by using FDM as
they cannot be solved using analytic methods. The DEs with FBCs is investigated as a
solution of a difference equation. Here, the FDM is applied to solve the second-order
DEs with FBCs for simplicity, but it is also applicable for higher-order DEs with
FBCs. So, in this way, our technique is reasonable and relevant for solving
higher-order DEs with FBCs.
CHAPTER 6
Methods for solving Higher Order
Differential Equations with Fuzzy
Initial and Boundary Conditions by
using Laplace Transforms
6.1 Introduction
Laplace transforms (LT) constitutes an important tool in solving linear ordinary and
partial DEs with constant coefficients and variable coefficients under suitable initial
and boundary conditions. The LT is a powerful integral transform used to switch a
function from the time domain to the s - domain, where s is a parameter. The Laplace
transform can be used in some cases to solve linear differential equations with given
initial conditions. Here, we use the Laplace transform technique mainly to solve
variable coefficient (not a Cauchy-Euler’s) DEs with FIBCs as well as forcing function
being a fuzzy number.
Opanuga et al. [61] used the Laplace transform method to solve BVPs. S.Salahshour
[71] focus on "nth order fuzzy linear DEs under GHD". Allahviranloo et al. [9]
propose a FLT under the SGHD concept; they use it in an analytic solution method for
some FDEs. Allahviranloo et al. [8] described "A new method for solving fuzzy linear
differential equations". Salahshour et al. [72] worked on "Application of fuzzy Laplace
Transform". Ahmad et al. [2] solved fuzzy two-point boundary value problem by FLT.
Salahshour et al. [73] solved the fuzzy heat equation using fuzzy Laplace transform.
85
Methods for solving Higher Order Differential Equations with Fuzzy Initial andBoundary Conditions by using Laplace Transforms 86
"Generalizations of the differentiability of fuzzy number valued functions with
applications to FDEs" were advocated by Bede et al. [18]. Hukurhara or its
generalization of Hukuhara derivatives, under this setting, existence and uniqueness of
solution of FDEs are studied. This approach has disadvantages that it gives the
solution which have increasing length of their support. The SGHD is defined for a
large class of fuzzy valued function than the HD, and FDEs can have solutions which
have a decreasing length of their support. But here, we have four possibilities for
selecting appropriate solution by using SGHD, and so, here we use strongly
generalized differentiability concept in this chapter.
6.2 Laplace transform
Let f (t) be a function defined, ∀ t ≥ 0. Then the integral, L [ f (t)] =∫
∞
0 f (t)e−stdt if
exists, is called LT of f (t). s is a parameter, real or complex number. Clearly L [ f (t)]
being a function of s is written as F(s) or L [ f (t)]. Here the symbol L which
transforms f (t) into F(s) is called the Laplace transform Operator.
6.2.1 Conditions for the existence of Laplace transforms
If f (t) is a piecewise continuous function on every finite interval in the range t ≥ 0
which satisfy | f (t)| ≤Meat , ∀ t ≥ 0 for some positive constants a and M, that is, f (t) is
of exponential order a, then the LT of f (t) i.e.∫
∞
0 f (t)e−stdt exists.
6.3 Strongly generalized Hukuhara differentiability
Let f : (a,b)→ E and t0 ∈ (a,b). We say that f is SGHD at t0 (Bede-Gal differential
[18]) if there exist an element f′(t0) ∈ E such that
Problem formulation 87
(i) for all h > 0 sufficiently small, ∃ f (t0 + h) f (t0), f (t0) f (t0− h) and the limits
(in the metric D)
limh→0
f (t0 +h) f (t0)h
= limh→0
f (t0) f (t0−h)h
= f′(t0).
or
(ii) for all h > 0 sufficiently small, ∃ f (t0) f (t0 +h), f (t0−h) f (t0) and the limits
(in the metric D)
limh→0
f (t0) f (t0 +h)−h
= limh→0
f (t0−h) f (t0)−h
= f′(t0).
or (iii) for all h > 0 sufficiently small, ∃ f (t0 + h) f (t0), f (t0− h) f (t0) and the
limits (in the metric D)
limh→0
f (t0 +h) f (t0)h
= limh→0
f (t0−h) f (t0)−h
= f′(t0).
or
(iv) for all h > 0 sufficiently small, ∃ f (t0) f (t0 +h), f (t0−h) f (t0) and the limits
(in the metric D)
limh→0
f (t0) f (t0 +h)−h
= limh→0
f (t0) f (t0−h)h
= f′(t0)
.
6.4 Problem formulation
Here we are going to discuss the solution of higher-order DEs with FIBCs by using
properties of LT. The LT is a versatile tool to solve higher order DEs because it converts
the DEs in form of algebraic equations and we know that the algebraic equations are
easy to deal with it. The other advantages are that we can solve the non-homogeneous
Methods for solving Higher Order Differential Equations with Fuzzy Initial andBoundary Conditions by using Laplace Transforms 88
DE directly without the need to first solving the corresponding homogeneous DE, or
find the particular solution of the DE without first finding the general solution. Here we
are going to solve some constant coefficient homogeneous and non-homogeneous DEs
with FICs. This method is mainly used to solve higher-order variable coefficient DEs
with FIBCs which are not of the Cauchy-Euler’s form of DEs.
6.5 Method for solving higher order differential
equations with fuzzy initial conditions
Fuzzy Laplace Transform(FLT)
Let f (t) be continuous fuzzy-valued function. Suppose that f (t)e−st improper FRI on
[0,∞). Then, the integral∫
∞
0 f (t)e−stdt is called FLT [9] and is defined as
L [ f (t)] =∫
∞
0 f (t)e−stdt (s > 0 and integer)
We have∫
∞
0 f (t)e−stdt = (∫
∞
0 f (t)e−stdt,∫
∞
0 f (t)e−stdt)
also by using definition of classical LT:
l[ f (t,r)] =∫
∞
0 f (t)e−stdt and
l[ f (t,r)] =∫
∞
0 f (t)e−stdt
then we follow
L [ f (t)] = (l[ f (t,r), l[ f (t,r)).
Theorem:6.5.1 Let f (t) and g(t) be continuous fuzzy-valued functions and c1,c2 are
real constants. Suppose that f (t)e−st ,g(t)e−st are improper FRI on [0,∞). Then
L [c1 f (t)+ c2g(t)] = c1L [ f (t)]+ c2L [g(t)].
Theorem:6.5.2 Let f : IR→ E be a function and denote f (t) = ( f (t,r), f (t,r)), for each
r ∈ [0,1]. Then
If f is (i)-differentiable, then f (t,r) and f (t,r) are differentiable function and f′(t) =
( f′(t,r), f
′(t,r)).
If f is (ii)-differentiable, then f (t,r) and f (t,r) are differentiable function and f′(t) =
( f′(t,r), f
′(t,r)).
Method for solving higher order differential equations with fuzzy initial conditions 89
Theorem:6.5.3 Let f′(t) be an integrable fuzzy-valued function [9], and f (t) is the
primitive of f′(t) on [0,∞). Then
L [ f′(t)] = sL [ f (t)] f (0)
where f is (i)-differentiable
or
L [ f′(t)] = (− f (0)) (−sL [ f (t)])
where f is (ii)-differentiable
Theorem:6.5.4 Let f′′(t) be integrable fuzzy-valued function [9], and f (t), f
′(t) are
primitive of f′(t), f
′′(t) on [0,∞). then
L [ f′′(t)] = s2L [ f (t)] s f (0) f
′(0)
where f is (i)-differentiable and f′is (i)-differentiable or
L [ f′′(t)] = s2L [ f (t)] s f (0)− f
′(0)
where f is (ii)-differentiable and f′is (ii)-differentiable or
L [ f′′(t)] =(−s2)L [ f (t)]− s f (0)− f
′(0)
where f is (i)-differentiable and f′is (ii)-differentiable or
L [ f′′(t)] =(−s2)L [ f (t)]− s f (0) f
′(0)
where f is (ii)-differentiable and f′is (i)-differentiable
Theorem:6.5.5 Let f (t) satisfy the condition of the existence theorem of LT and
L [ f (t)] = F(s), then L [t f (t)] =−F′(s)
Hence if f′(t) satisfies the condition of existence theorem of LT then
L [t f′(t)] =− d
dsL [ f
′(t)] =− d
ds
{sF(s)− f (0)
}=−sF
′(s)− F(s)
Similarly, for f′′
L [t f′′(t)]=− d
dsL [ f
′′(t)]=− d
ds
{s2F(s)−s f (0)− f
′(0)}=−s2F
′(s)−2sF(s)+ f (0)
Methods for solving Higher Order Differential Equations with Fuzzy Initial andBoundary Conditions by using Laplace Transforms 90
6.6 Solution of higher order differential equations with
fuzzy initial and boundary conditions
6.6.1 Solution of second order constant coefficient differential
equations with fuzzy initial conditions
Application of simple harmonic motion in different fields
• In the simple harmonic motion, the displacement of the object is always in the
opposite direction of the restoring force.
• Simple harmonic motion is always oscillatory.
• Examples are: the motion of a pendulum, motion of a spring, etc.
• Its applications can be found in clock, guitar, violin, bungee jumping, rubber
bands, diving boards, earthquakes, etc.
6.6.1.1 An example based on the application of differential equation in
mechanical engineering
Here, we consider the motion of an object with mass at the end of a spring that is either
vertical or horizontal on a level surface. By Hooke’s Law, which states that if the spring
is stretched (or compressed) units from its natural length, then it exerts a force that is
proportional to y,
restoring force = ky
where k is a positive constant (called the spring constant). Besides, consider the
motion of a spring that is subject to a frictional force (due to air resistance or friction).
The damping force is proportional to the velocity of the mass and acts in the direction
opposite to the motion.
Solution of higher order differential equations with fuzzy initial and boundaryconditions 91
damping force = by′
where b is the damping coefficient.
By using Newton’s Second Law (force equals mass times acceleration), we have
my′′
= restoring force + damping force = −ky−by′
By using above equations, we get the equation of mass-spring system (vibration of
mass m on an spring) and it is modelled by the homogeneous linear ODE
my′′+by
′+ ky = 0 (6.1)
Here y(t), a function of time t, is the displacement of body of mass m from rest. There
are free motions, that is, motions in the absence of external forces.
Consider non-homogeneous linear ODE
my′′+by
′+ ky = r(t) (6.2)
where r(t) is a external force and the resultant motion is the forced motion.
m
y(t)
kr(t)
b
FIGURE 6.1: The mass-spring system
Example:6.1 Consider vibrating mass system. The mass m = 1, the spring constant
k = 4 lbs/ft and there is no or negligible damping. The forcing function is 2cos t.
Here b = 0. So, from (6.2) the DE of the motion with initial conditions is
Methods for solving Higher Order Differential Equations with Fuzzy Initial andBoundary Conditions by using Laplace Transforms 92
y′′+4y = 2cos t, t ∈ [0,1]
y(0) = 2
y′(0) = 0
Consider DE with FICs
y′′+4y = 2cos t, t ∈ [0,1]
y(0) = [2r, 4−2r]
y′(0) = [−2+2r, 2−2r]
Apply LT on both the side
L [y′′]+4L [y] = 2L [cos t]
By using Theorems 6.5.1 and 6.5.4
l[y(t,r)] = (2r)s
s2 +4+(2r−2)
1s2 +4
+2s
(s2 +1)(s2 +4)
l[y(t,r)] = (4−2r)s
s2 +4+(2−2r)
1s2 +4
+2s
(s2 +1)(s2 +4)Hence the solution bounds are:
y(t,r) = (2r)cos2t + r sin2t− sin2t +23
cos t− 23
cos2t
y(t,r) = (4−2r)cos2t + sin2t− r sin2t +23
cos t− 23
cos2t
The y(t,r) and y(t,r) values at r = 0,0.5,0.8,0.9 are represented in Figs.6.2 to 6.5.
when r = 1, we obtain exact solution
y(t) = y(t,1) = y(t,1) =43
cos2t +23
cos t,
as shown in Fig.6.6.
0.2 0.4 0.6 0.8 1.0t
1
2
3
4
y
FIGURE 6.2: The lower-bound and upper-bound of solution at r = 0
Solution of higher order differential equations with fuzzy initial and boundaryconditions 93
0.2 0.4 0.6 0.8 1.0t
0.5
1.0
1.5
2.0
2.5
3.0
y
FIGURE 6.3: The lower-bound and upper-bound of solution at r = 0.5
0.2 0.4 0.6 0.8 1.0t
0.5
1.0
1.5
2.0
y
FIGURE 6.4: The lower-bound and upper-bound of solution at r = 0.8
0.2 0.4 0.6 0.8 1.0t
0.5
1.0
1.5
2.0
y
FIGURE 6.5: The lower-bound and upper-bound of solution at r = 0.9
0.2 0.4 0.6 0.8 1.0t
0.5
1.0
1.5
2.0
y
FIGURE 6.6: The lower-bound and upper-bound of solution at r = 1
Methods for solving Higher Order Differential Equations with Fuzzy Initial andBoundary Conditions by using Laplace Transforms 94
6.6.1.2 Example based on the application of differential equation in physics
An ideal simple pendulum consists of a point mass m1 suspended from a support by a
massless string of length l. The equilibrium position of the mass is a distance l below
the support. If the mass is displaced from its equilibrium position, it exhibits periodic
motion, moving in a vertical plane along a circular arc. When the string makes an
angle θ with the vertical, then the displacement of the mass from its equilibrium
position along the circular arc is s = lθ . The forces acting on the mass are gravity and
the tension in the string. Only gravity provides a restoring force towards the
equilibrium position. The magnitude of this force is m1 gsinθ . The equation of motion,
F = m1a, therefore yields
restoring force = m1d2sdt2 =−m1 gsinθ ,
or
restoring force = ld2θ
dt2 =−gsinθ .
Here, the displacement from equilibrium is small ⇒ sinθ ∼ θ in radians.(Fig.6.7)
Then the equation of motion becomes
ld2θ
dt2 =−gθ .
And, the resistive force is proportional to velocity
The resistive force = − lλm1
dθ
dt.
Solution of higher order differential equations with fuzzy initial and boundaryconditions 95
Therefore,
ld2θ
dt2 = resistive force + restoring force = − lλm1
dθ
dt−gθ
A simple pendulum of length l oscillating through a small angle θ in a medium
resistance is proportional to velocity. The differential equation of motion is
ld2θ
dt2 +lλm1
dθ
dt+gθ = 0 (6.3)
m1
1g
θ
FIGURE 6.7: The simple pendulum
Example:6.2 A Pendulum of length l = 85 f t is subject to resistive force FR = 32
5dθ
dt due
to damping. Determine displacement function.
Equation (6.3) gives the differential equation as below
85
d2θ
dt2 +325
dθ
dt+32θ = 0
θ(0) = 1
θ′(0) = 2
Methods for solving Higher Order Differential Equations with Fuzzy Initial andBoundary Conditions by using Laplace Transforms 96
By simplifying and converting them to DE with FICs, The DE become
θ′′+4θ
′+20θ = 0
θ(0) = [r,2− r]
θ′(0) = [1+ r,3− r]
Apply LT on both the side
L [θ′′]+4L [θ ]+20L [θ ] = L [0]
By using Theorems 6.5.1 to 6.5.4
l[θ(t,r)] = rs+2
(s+2)2 +16+
(3r+1
4
)1
(s+2)2 +16
l[θ(t,r)] = (2− r)s+2
(s+2)2 +16+
(7−3r
4
)1
(s+2)2 +16Hence the solution bounds are as follows:
θ(t,r) = re−2t cos4t +(
3r+14
)e−2t sin4t
θ(t,r) = (2− r)e−2t cos4t +(
7−3r4
)e−2t sin4t
The θ(t,r) and θ(t,r) values at r = 0,0.5,0.8,0.9 are represented in Figs.6.8 to 6.11.
when r = 1, we obtained exact solution
θ(t) = θ(t,1) = θ(t,1) = e−2t cos4t + e−2t sin4t,
as shown in Fig.6.12.
0.2 0.4 0.6 0.8 1.0t
0.5
1.0
1.5
2.0
Θ
FIGURE 6.8: The lower-bound and upper-bound of solution at r = 0
Solution of higher order differential equations with fuzzy initial and boundaryconditions 97
0.2 0.4 0.6 0.8 1.0t
0.5
1.0
1.5
Θ
FIGURE 6.9: The lower-bound and upper-bound of solution at r = 0.5
0.2 0.4 0.6 0.8 1.0t
0.5
1.0
Θ
FIGURE 6.10: The lower-bound and upper-bound of solution at r = 0.8
0.2 0.4 0.6 0.8 1.0t
-0.2
0.2
0.4
0.6
0.8
1.0
1.2
Θ
FIGURE 6.11: The lower-bound and upper-bound of solution at r = 0.9
0.2 0.4 0.6 0.8 1.0t
-0.2
0.2
0.4
0.6
0.8
1.0
Θ
FIGURE 6.12: The lower-bound and upper-bound of solution at r = 1
Methods for solving Higher Order Differential Equations with Fuzzy Initial andBoundary Conditions by using Laplace Transforms 98
6.6.2 Solution of second order variable coefficient differential
equations with fuzzy initial and boundary conditions
Example:6.3 Consider variable coefficient DE with initial condition
ty′′− y
′=−1, t ∈ [0,1]
y(0) = 0
Consider variable coefficient DE with initial condition and forcing function both as
TFN
ty′′− y
′=−1, t ∈ [0,1]
y(0) = [r−1,1− r]
1 = [r,2− r]
Apply LT on both the side
L [ty′′]−L [y
′] =−L [1]
By using Theorems 6.5.1 to 6.5.5
l[y(t,r)] =r−1
s+
rs2 +
As3
l[y(t,r)] =1− r
s+
2− rs2 +
As3 , A ∈ R.
Hence the solution bounds are as follows:
y(t,r) = (r−1)+ rt +Bt2
y(t,r) = (1− r)+(2− r)t +Bt2, B ∈ R.
The y(t,r) and y(t,r), 0≤ r ≤ 1, are represented in Fig.6.13
y(t) = y(t,1) = y(t,1) = t +Bt2
0.00
0.05
0.10
t
0.0
0.5
1.0
r
-1.0
-0.5
0.0
0.5
1.0
yHt,rL
FIGURE 6.13: Solution by considering parameters TFN
Solution of higher order differential equations with fuzzy initial and boundaryconditions 99
Consider initial condition and forcing function as TraFN
ty′′− y
′=−1, t ∈ [0,1]
y(0) = [r−1,2− r]
1 = [r,3− r]
Apply LT on both the side
L [ty′′]−L [y
′] =−L [1]
By using Theorems 6.5.1 to 6.5.5
l[y(t,r)] =r−1
s+
rs2 +
As3
l[y(t,r)] =2− r
s+
3− rs2 +
As3 , A ∈ R.
Hence the solution bounds are as follows:
y(t,r) = (r−1)+ rt +Bt2
y(t,r) = (2− r)+(3− r)t +Bt2, B ∈ R.
The y(t,r) and y(t,r), 0≤ r ≤ 1, are represented in Fig.6.14
y(t,1) = t +Bt2 , y(t,1) = 1+2t +Bt2
0.00
0.05
0.10
t
0.0
0.5
1.0
r
-1
0
1
2
yHt,rL
FIGURE 6.14: Solution by considering parameters TraFN
Consider initial condition and forcing function as GFN
ty′′− y
′=−1, t ∈ [0,1]
y(0) = [−√−(2loger),
√−(2loger)]
1 = [1−√(−2loger),1+
√(−2loger)]
Apply LT on both the side
L [ty′′]−L [y
′] =−L [1]
Methods for solving Higher Order Differential Equations with Fuzzy Initial andBoundary Conditions by using Laplace Transforms 100
By using Theorems 6.5.1 to 6.5.5
l[y(t,r)] =−√−(2loger)
s+
1−√
(−2loger)s2 +
As3
l[y(t,r)] =
√−(2loger)
s+
1+√
(−2loger)s2 +
As3 , A ∈ R.
Hence the solution bounds are as follows:
y(t,r) = [−√−(2loger)]+ [1−
√(−2loger)]t +Bt2
y(t,r) = [√−(2loger)]+ [1+
√(−2loger)]t +Bt2, B ∈ R.
The y(t,r) and y(t,r), 0≤ r ≤ 1, are represented in Fig.6.15
y(t) = y(t,1) = y(t,1) = t +Bt2
0.00
0.05
0.10
t
0.0
0.5
1.0
r
-5
0
5
10
yHt,rL
FIGURE 6.15: Solution by considering parameters GFN
Example:6.4 Consider variable coefficient DE with boundary conditions
ty′′+2y
′+ ty = 0,
y(0) = 1
y(π) = 0
Consider variable coefficient DE with boundary conditions as TFN
ty′′+2y
′+ ty = 0,
y(0) = [r,2− r]
y(π) = [r−1,1− r]
Apply LT on both the side
L [ty′′]+2L [y
′]+L [ty] = L [0]
Solution of higher order differential equations with fuzzy initial and boundaryconditions 101
By using Theorems 6.5.1 to 6.5.5
l[y(t,r)] = (r) tan−1(1s )
l[y(t,r)] = (2− r) tan−1(1s )
Hence the solution bounds are as follows:
y(t,r) = (r)sin t
ty(t,r) = (2− r)
sin tt
The y(t,r) and y(t,r), 0≤ r ≤ 1, are represented in Fig.6.16
y(t) = y(t,1) = y(t,1) =sin t
t
0.0
0.5
1.0r
01 2
3t
0.0
0.5
1.0
1.5
2.0
yHt,rL
FIGURE 6.16: Solution by considering parameters TFN
Consider boundary conditions as TraFN
ty′′+2y
′+ ty = 0,
y(0) = [r,3− r]
y(π) = [r−1,2− r]
Apply LT on both the side
L [ty′′]+2L [y
′]+L [ty] = L [0]
By using Theorems 6.5.1 to 6.5.5
l[y(t,r)] = (r) tan−1(1s )
l[y(t,r)] = (3− r) tan−1(1s )
Hence the solution bounds are as follows:
y(t,r) = (r)sin t
ty(t,r) = (3− r)
sin tt
The y(t,r) and y(t,r), 0≤ r ≤ 1, are represented in Fig.6.17
y(t,1) =sin t
t, y(t,1) = 2
sin tt
Methods for solving Higher Order Differential Equations with Fuzzy Initial andBoundary Conditions by using Laplace Transforms 102
0.0
0.5
1.0r
01 2
3
t
0
1
2
3
yHt,rL
FIGURE 6.17: Solution by considering parameters TraFN
Consider boundary conditions as GFN
ty′′+2y
′+ ty = 0,
y(0) = [1−√−(2loger),1+
√−(2loger)]
y(π) = [−√(−2loger),
√(−2loger)]
Apply LT on both the side
L [ty′′]+2L [y
′]+L [ty] = L [0]
By using Theorems 6.5.1 to 6.5.5
l[y(t,r)] = [1−√
(−2loger)] tan−1(1s )
l[y(t,r)] = [1+√
(−2loger)] tan−1(1s )
Hence the solution bounds are as follows:
y(t,r) = [1−√
(−2loger)]sin t
ty(t,r) = [1+
√(−2loger)]
sin tt
The y(t,r) and y(t,r) , 0≤ r ≤ 1, are represented in Fig.6.18
y(t) = y(t,1) = y(t,1) =sin t
t
0.0
0.5
1.0r
01 2
3
t
0
5
yHt,rL
FIGURE 6.18: Solution by considering parameters GFN
Result and Discussion 103
6.7 Result and Discussion
The Laplace transform method is mainly used to solve variable coefficient (not a
Cauchy-Euler’s) DEs with FIBCs. Here, we use different types of fuzzy numbers as
fuzzy conditions to see the change in the solution. The selection of a fuzzy number
depends on situation and the type of the problem. The graphical representation is done
using the Mathematica software. From the figures, we can observe that for distinct
values of r, we get different lower and upper bounds, whereas at r = 1 the lower and
the upper bound are nearly the same. By using the LT technique we have solved the
DEs with FBCs and second-order DE in which a single condition is given.
6.8 Conclusion
The LT technique offers us a solution of higher-order DEs with FICs but here for the
simplicity purpose we have considered second order constant coefficient and variable
coefficient DEs with FIBCs by using the SGHD notion. The usefulness of the method
was illustrated by solving some examples, including application-level examples. The
uniqueness of the solution is lost in this new procedure because we have four
possibilities and the solution should be appropriately selected among the four cases of
SGHD, but such situation in a fuzzy context is assumed. Here we solved DEs with
FIBCs by using the fuzzy numbers like TFN, TraFN and GFN.
References
[1] O Abu-Arqub, A El-Ajou, S Momani and N Shawagfeh, Analytical solutions of fuzzy initial valueproblems by HAM, Applied Mathematics & Information Sciences, 7(5), (2013), 1903.
[2] L Ahmad, M Farooq, S Ullah and S Abdullah, Solving fuzzy two-point boundary value problemusing fuzzy Laplace transform, arXiv preprint arXiv:1403.0571.
[3] M Z Ahmad and M K Hasan, A new fuzzy version of Euler’s method for solving differentialequations with fuzzy initial values, Sains Malaysiana, 40(6), (2011), 651–657.
[4] M Z Ahmad, M K Hasan and B D Baets, Analytic and numerical solutions of fuzzy differentialequations, Information sciences, 236(1), (2013), 156–161.
[5] M B Ahmadi, N A Kiani and N Mikaeilvand, Laplace transform formula on fuzzy nth-orderderivative and its application in fuzzy ordinary differential equations, Soft Computing, 18(12),(2014), 2461–2469.
[6] Ö Akin, T Khaniyev, Ö Oruç and I B TüRksEn, An algorithm for the solution of second order fuzzyinitial value problems, Expert Systems with Applications, 40(3), (2013), 953–957.
[7] T Allahviranloo, S Abbasbandy, N Ahmady and E Ahmady, Improved predictor-corrector methodfor solving fuzzy initial value problems, Information Sciences, 179(7), , 945–955.
[8] T Allahviranloo, S Abbasbandy, S Salahshour and A Hakimzadeh, A new method for solving fuzzylinear differential equations, Computing, 92(2), (2011), 181–197.
[9] T Allahviranloo and M B Ahmadi, Fuzzy laplace transforms, Soft Computing, 14(3), (2010), 235–243.
[10] T Allahviranloo, E Ahmady and N Ahmady, Nth-order fuzzy linear differential equations,Information sciences, 178(5), (2008), 1309–1324.
[11] T Allahviranloo, E Ahmady and N Ahmady, A method for solving n-th order fuzzy lineardifferential equations, International Journal of Computer Mathematics, 86(4), (2009), 730–742.
[12] T Allahviranloo, N A Kiani and M Barkhordari, Toward the existence and uniqueness of solutionsof second-order fuzzy differential equations, Information sciences, 179(8), (2009), 1207–1215.
[13] A Armand and Z Gouyandeh, Solving two-point fuzzy boundary value problem usingthe variational iteration method, Communications on Advanced Computational Science withApplications, 2013, (2013), 1–10.
[14] J P Aubin and A Cellina, Theory of fuzzy differential equations and inclusions (2006).
[15] F Bahrami, K Ivaz and A Khastan, New result on multiple solutions for Nth-order fuzzy differentialequations under generalized differentiability, Boundary Value Problems, 395714(1), (2009), 1–13.
[16] B Bede, Note on "Numerical solutions of fuzzy differential equations by predictor–correctormethod”, Information sciences, 178(7), (2008), 1917–1922.
[17] B Bede, T G Bhaskar and V Lakshmikantham, Perspectives of fuzzy initial value problems,Communications in Applied Analysis, 11(3-4), (2007), 339–358.
[18] B Bede and S G Gal, Generalizations of the differentiability of fuzzy-number-valued functions withapplications to fuzzy differential equations, Fuzzy sets and systems, 151(3), (2005), 581–599.
104
References 105
[19] B Bede and S G Gal, Remark on the new solutions of fuzzy differential equations, Chaos SolitonsFractals.
[20] B Bede and S G Gal, Solutions of fuzzy differential equations based on generalized differentiability,Communications in Mathematical Analysis, 9(2), (2010), 22–41.
[21] B Bede, I J Rudas and A L Bencsik, First order linear fuzzy differential equations under generalizeddifferentiability, Information sciences, 177(7), (2007), 1648–1662.
[22] J J Buckley and T Feuring, Fuzzy initial value problem for Nth-order linear differential equations,Fuzzy Sets and Systems, 121(2), (2001), 247–255.
[23] S Chakraverty, S Tapaswini and D Behera, Fuzzy Differential Equations and Applications forEngineers and Scientists, CRC Press (2016).
[24] Y Chalco-Cano and H Roman-Flores, On new solutions of fuzzy differential equations, Chaos,Solitons & Fractals, 38(1), (2008), 112–119.
[25] S L Chang and L A Zadeh, On fuzzy mapping and control, IEEE Transaction on System ManCybernet, 2, (1972), 30–34.
[26] S L Chang and L A Zadeh, Fuzzy sets, fuzzy logic, and fuzzy systems, World Scientific (1996).
[27] M Chen, Y Fu, X Xue and C Wu, Two-point boundary value problems of undamped uncertaindynamical systems, Fuzzy Sets and Systems, 159(16), (2008), 2077–2089.
[28] M Chen, C Wu, X Xue and G Liu, On fuzzy boundary value problems, Information Sciences,178(7), (2008), 1877–1892.
[29] N B Desai, Fuzzy Modeling and its Applications, M.phil Dessertation, S.G.University,Surat (1997).
[30] Z Ding, M Ma and A Kandel, Existence of the solutions of fuzzy differential equations withparameters, Information Sciences, 99(3), (1997), 205–217.
[31] D Dubois and H Prade, Towards fuzzy differential calculus part 3: Differentiation, Fuzzy sets andsystems, 8(3), (1982), 225–233.
[32] A Esfahani, O S Fard and T A Bidgoli, On the existence and uniqueness of solutions to fuzzyboundary value problems, Ann. Fuzzy Math. Inform, 7(1), (2014), 15–29.
[33] M Friedman, M Ming and A Kandel, Numerical solutions of fuzzy differential and integralequations, Fuzzy sets and Systems, 106(1), (1999), 35–48.
[34] N Gasilov, E Amrahov and A G Fatullayev, Linear differential equations with fuzzy boundaryvalues, in: 2011 5th International Conference on Application of Information and CommunicationTechnologies (AICT), IEEE (2011), 1–5.
[35] N A Gasilov, E Amrahov and A G Fatullayev, Solution of linear differential equations with fuzzyboundary values, Fuzzy Sets and Systems, 257, (2014), 169–183.
[36] N A Gasilov, S Amrahov, A G Fatullayev and I F Hashimoglu, Solution method for a boundaryvalue problem with fuzzy forcing function, Information Sciences, 317, (2015), 349–368.
[37] N A Gasilov, A G Fatullayev, E Amrahov and A Khastan, A new approach to fuzzy initial valueproblem, Soft Computing, 18(2), (2014), 217–225.
[38] N A Gasilov, I F Hashimoglu, S E Amrahov and A G Fatullayev, A new approach to non-homogeneous fuzzy initial value problem, Computer Modeling in Engineering & Sciences(CMES),85(4), (2012), 367–378.
References 106
[39] D N Georgiou, J J Nieto and R Rodriguez-Lopez, Initial value problems for higher-order fuzzydifferential equations, Nonlinear Analysis: Theory, Methods & Applications, 63(4), (2005), 587–600.
[40] R Goetschel and W Voxman, Elementary fuzzy calculus, Fuzzy sets and systems, 18(1), (1986),31–43.
[41] L T Gomes, de Barros, C Laécio and B Bede, Fuzzy differential equations in various approaches,Springer (2015).
[42] K Ivaz, A Khastan and J J Nieto, A numerical method for fuzzy differential equations and hybridfuzzy differential equations, in: Abstract and Applied Analysis, Hindawi (2013).
[43] H Jafari, S Mohammad and D Baleanu, The variational iteration method for solving n-th order fuzzydifferential equations, Open Physics, 10(1), (2012), 76–85.
[44] O Kaleva, Fuzzy differential equations, Fuzzy sets and systems, 24(3), (1987), 301–317.
[45] O Kaleva, The Cauchy problem for fuzzy differential equations, Fuzzy sets and systems, 35(3),(1990), 389–396.
[46] O Kaleva, A note on fuzzy differential equations, Nonlinear Analysis: Theory, Methods &Applications, 64(5), (2006), 895–900.
[47] A Kandel, Fuzzy dynamical systems and the nature of their solutions, Springer (1980).
[48] A Kandel and W J Byatt, Fuzzy differential equations, Proceeding of the international conferenceon cybernetics and society (1978).
[49] A Khastan, J Nieto and R Rodríguez-López, Variation of constant formula for first order fuzzydifferential equations, Fuzzy Sets and Systems, 177, (2011), 20–33.
[50] A Khastan and J J Nieto, A boundary value problem for second order fuzzy differential equations,Nonlinear Analysis: Theory, Methods & Applications, 72(9), (2010), 3583–3593.
[51] G J Klir and B Yuan, Fuzzy sets and Fuzzy logic Theory and Applications, Prentice Hall (1995).
[52] V Lakshmikantham and R N Mohapatra, Theory of fuzzy differential equations and inclusions,CRC Press (2004).
[53] V Lakshmikantham, K Murty and J Turner, Two-point boundary value problems associated withnon-linear fuzzy differential equations, Mathematical Inequalities and Applications, 4, (2001),527–534.
[54] V Lakshmikantham and J J Nieto, Differential equations in metric spaces: an introduction andan application to fuzzy differential equations, DYNAMICS OF CONTINUOUS DISCRETE ANDIMPULSIVE SYSTEMS SERIES A, 10, (2003), 991–1000.
[55] V Laksmikantham, S Leela and A Vatsala, Interconnection between set and fuzzy differentialequations, Nonlinear Analysis: Theory, Methods & Applications, 54(2), (2003), 351–360.
[56] M Ma, M Friedman and A Kandel, Numerical solutions of fuzzy differential equations, Fuzzy setsand systems, 105(1), (1999), 133–138.
[57] S Mansouri and N Ahmady, A numerical method for solving nth-order fuzzy differential equationby using characterization theorem, Communication in Numerical Analysis, (2012), 12.
[58] S P Mondal and T P Roy, First order linear non homogeneous ordinary differential equation in fuzzyenvironment based on Laplace transform, Journal of Mathematical and Computational Science,3(6), (2013), 1533–1564.
References 107
[59] J J Nieto, The Cauchy problem for continuous fuzzy differential equations, Fuzzy sets and systems,102(2), (1999), 259–262.
[60] J J Nieto, A Khastan and K Ivaz, Numerical solution of fuzzy differential equations undergeneralized differentiability, Nonlinear Analysis: Hybrid Systems, 3(4), (2009), 700–707.
[61] A A Opanuga, E A Owoloko, H I Okagbue and O O Agboola, Finite difference method and Laplacetransform for boundary value problems.
[62] D O’Regan, V Lakshmikantham and J J Nieto, Initial and boundary value problems for fuzzydifferential equations, Nonlinear Analysis: Theory, Methods & Applications, 54(3), (2003), 405–415.
[63] S C Palligkinis, G Papageorgiou and I T Famelis, Runge-Kutta methods for fuzzy differentialequations, Applied Mathematics and Computation, 209(1), (2009), 97–105.
[64] K R Patel and N B Desai, Solution of fuzzy initial value problems, ADIT Journal of Engineering,12(1), (2015), 53–57.
[65] K R Patel and N B Desai, Solution of fuzzy initial value problem by fuzzy Laplace transform, vol. 2,Kalpa Publications in Computing, ICRISET2017, Selected Papers in Computing (2017), 25–37.
[66] K R Patel and N B Desai, Solution of nth order fuzzy initial value problem, PRAJNA- Journal ofPure and Applied Sciences, (24-25), (2017), 30–37.
[67] K R Patel and N B Desai, Solution of variable co-efficient fuzzy differential equations byfuzzy Lapalce transform, International journal recdent and innovative trends in computing andcommunication, 5(6), (2017), 927–942.
[68] K R Patel and N B Desai, Solution of higher order differential equations with fuzzy boundaryconditions by Finite difference method, Journal of Emerging Technologies and InnovativeResearch, 6(5), (2019), 30–49.
[69] M L Puri and D A Ralescu, Differentials of fuzzy functions, Journal of Mathematical Analysis andApplications, 91(2), (1983), 552–558.
[70] T J Ross, Fuzzy Logic with Engineering Applications, Wiley (2013).
[71] S Salahshour, Nth-order fuzzy differential equations under generalized differentiability, Journal ofFuzzy Set Valued Analysis, 2011, (2011), 1–14.
[72] S Salahshour and T Allahviranloo, Applications of fuzzy Laplace transforms, Soft computing, 17(1),(2013), 145–158.
[73] S Salahshour and E Haghi, Solving fuzzy heat equation by fuzzy Laplace transforms,in: International Conference on Information Processing and Management of Uncertainty inKnowledge-Based Systems, Springer (2010), 512–521.
[74] S Seikkala, On the fuzzy initial value problems, Fuzzy Sets and Systems, 24(3), (1987), 319–330.
[75] S Song and C Wu, Existence and uniqueness of solutions to Cauchy problem of fuzzy differentialequations, Fuzzy sets and Systems, 110(1), (2000), 55–67.
[76] S Tapaswini and S Chakraverty, Numerical solution of n-th order fuzzy linear differential equationsby homotopy perturbation method, International Journal of Computer Applications, 64(6).
[77] H C Wu, The fuzzy Riemann integral and its numerical integration, Fuzzy Sets and Systems, 110(1),(2000), 1–25.
[78] G Xiaobin, S Dequan and L Xiaoquan, Fuzzy approximate solutions of second-order fuzzy linearboundary value problems, Boundary Value Problems, 2013(1), (2013), 212.
References 108
[79] B Yuan et al., Fuzzy Sets, Fuzzy Logic, and Fuzzy Systems: Selected Papers by Lotfi A Zadeh,World Scientific (1996).
[80] L A Zadeh, Fuzzy sets, Information and control, 8(3), (1965), 338–353.
[81] L A Zadeh, The concept of a linguistic variable and its application to approximate reasoning—I,Information sciences, 8(3), (1975), 199–249.
[82] H J Zimmermann, Fuzzy set theory and its applications, Springer Science & Business Media (2011).
List of Publications
List of Publications Arising From the Thesis
1. Solution of fuzzy initial value problems, ADIT Journal of Engineering, ISSN Number : 0973-
3663,(UGC approved journal), 12(1), (2015), 53-58.
2. Solution of variable coefficient fuzzy differential equations by fuzzy Laplace transform,
International Journal on Recent and Innovation Trends in Computing and Communication,
ISSN Number : 2321-8169,(UGC approved journal), 5(6), (2017), 927-942.
3. Solution of nth order fuzzy initial value problem, PRAJNA- Journal of Pure and Applied
Sciences,ISSN Number : 0975-2595,(UGC approved journal) 24-25,(2017), 30-37.
4. Solution of higher order differential equations with fuzzy boundary conditions by finite difference
method, Journal of Emerging Technologies and Innovative Research, ISSN Number : 2349-
5162,(UGC approved journal), 6(5), (2019), 30-49.
Details of the Work Presented in Conference From the Thesis
1. The paper entitled as "Solution of fuzzy initial value problems by fuzzy Laplace transform"
presented in ICRISET2017 at B. V. M. Engineering College, V. V. Nagar, Anand (Gujarat) on
18th February, 2017 and published in Kalpa Publications in Computing, ICRISET2017, Selected
Papers in Computing, 2, (2017), 25-37.
Details of the Work Accepted From the Thesis
1. The paper entitled as "Solution of nth Order Differential Equation with Fuzzy Conditions by
Gauss Elimination Method" is accepted by the "Continuity, Consistency and Innovation in
Applied Sciences and Humanities" on August 13, 2020.
109
Solution of Variable Coefficient Fuzzy
Differential Equations by Fuzzy Laplace
Transform
Komal R.PatelNarendrasinh B.Desai
Assistant ProfessorInstitute of Technology and Management Universe
Vadodara-390510,Gujarat(INDIA)
Associate ProfessorHead of Department of Applied Sciences and Humanities
A. D. Patel Institute of TechnologyV. V. Nagar-388121,Gujarat(INDIA)
Email:[email protected]@yahoo.co.in
Abstract
In this paper we propose a fuzzy Laplace transform to solve variablecoefficient fuzzy differential equations under strongly generalized differ-entiability concept.The fuzzy Laplace transform of derivative was usedto solve second order variable coefficient fuzzy initial value problems andfuzzy boundary value problems if t is multiplied with first or second deriva-tive term.To illustrate applicability of proposed method we solve fuzzydifferential equations using different types of fuzzy numbers i.e. triangu-lar,trapezoidal,Gaussian etc and compare the solutions.We plot 3D plotsfor different values of r-level sets by mathematica software.
Keywords:Fuzzy Number.Fuzzy valued function.triangular,trapezoidal and Gaus-sian fuzzy numbers.Fuzzy Laplace Transform.Strongly generalized differential.Fuzzyinitial value problem.Fuzzy boundary value problem..
1 Introduction
The fuzzy differential equation is very much important topic in field of scienceand engineering to solve dynamic problem.The concept of a fuzzy derivativewas first introduced by Chang and Zadeh [56],followed up by Dubois and Prade
1
Publication 1 110
[17] who used the extension principle in their approach.Other fuzzy derivativeconcepts were proposed by Puri and Ralescu [45], and Goetschel and Vaxman[26]as an extension of the Hukuhara derivative of multivalued functions.Kandeland Byatt [33] applied the concept of fuzzy differential equation to the analysisof fuzzy dynamical problems.The fuzzy differential equations and fuzzy initialvalue problems are studied by Kaleva[31, 32]and Seikkala [51]
Two analytical methods for solving an nth-order fuzzy linear differentialequation with fuzzy initial conditions presented by Buckley and Feuring [12,13].Mondal and Roy [42] described the solution procedure for first order linearnon-homogeneous ordinary differential equation in fuzzy environment.Existenceand uniqueness of fuzzy boundary value problem has been proved by Esfahani etal.[18].Lakshikantham et al. [38] investigated the solution of two point boundaryvalue problems associated with non-linear fuzzy differential equation by usingthe extension principle.Generalized differentiability concept is used by Bade etal.[11] to investigated first order linear fuzzy differential equations.Based on theidea of collocation method Allahviranloo et al. [5] solved nth order fuzzy lin-ear differential equations.Far and Ghal-Eh [19] proposed an iterative methodto solve fuzzy differential equations for the linear system of first order fuzzydifferential equation with fuzzy constant coefficient.Variation of constant for-mula has been handle by Khastan et al.[37] to solve first order fuzzy differentialequations.Akin et al.[2] developed an algorithm based on α-cut of fuzzy set forsolution of second order fuzzy initial value problems.A new approach has beendeveloped by Gasilov et al.[22] to get the solution of fuzzy initial value problem.
The concept of generalized H-differentiability is studied by Chalco-Cano andRoman Flores [14] to solve fuzzy differential equation.Hasheni et al.[29, 28] stud-ied homotopy analysis method for solution of system of fuzzy differential equa-tion s and obtained analytic solution of fuzzy Wave like equations with variablecoefficients.As regards, methods to solve nth order fuzzy differential equationare discussed in [5, 25, 30, 35, 48, 55].the Variational iteration method (VIM)was successfully applied by Jafari et al.[30] for solving nth order fuzzy differen-tial equation.A new result on multiple solutions for nth order fuzzy differentialequations under generalized differentiability has been proposed by Khastan etal.[35].Based on idea of collocation method allahviranloo et al.[5] solved nth or-der fuzzy linear differential equations.Integral form of nth order fuzzy differentialequations has been developed by Salahshour [48]under generalized differentiabil-ity.Mansuri and Ahmady [41] implemented characterization theorem for solvingnth order fuzzy differential equations.Also Tapaswini and Chakraverty[53] im-plemented homotopy perturbation method for the solution of nth order fuzzylinear differential equations.Bade[10] found Solutions of fuzzy differential equa-tions based on generalized differentiability.
Paper is organized as In section 2 preliminaries,In section 3 Examples byusing fuzzy Laplace transform ,In section 4 Result and Discussion,In section 5conclusion.
Publication 1 111
2 Preliminaries
Definition 2.1 Fuzzy NumberA fuzzy number is a fuzzy set like µ : R→ I = [0, 1] which satisfies:
(a) µ is upper semi-continuous,
(b) µ is fuzzy convex i.e µ(λx+(1−λ)y) ≥ min{µ(x), µ(y)}∀x, y ∈ R, λ ∈ [0, 1],
(c) µ is normal i.e ∃x0 ∈ R for which µ(x0) = 1,
(d) supp µ = {x ∈ R | µ(x) > 0} is support of u, and its closure cl(supp µ) iscompact.
Definition 2.2 r-cutIt is crisp set derived from its parent fuzzy set A where r-cut is defined asAr = {x ∈ R | µ(x) ≥ r}Definition 2.3 Triangular Fuzzy NumberConsider triangular fuzzy number A = (a, b, c) is depicted in Fig.1 The mem-bership function µ(x) of A will be defined as follows.
µ(x) =
0 , x < ax−ab−a , a ≤ x ≤ bc−xc−b , b ≤ x ≤ c0 , x > c
The triangular fuzzy number A = (a, b, c) can be represented with an order pairof function of r-cut approach i.e.[µ(r), µ(r)] = [a+ (b− a)r, c− (c− b)r] , wherer ∈ [0, 1]
Fig.1 Triangle membership function
Definition 2.4 Trapezoidal Fuzzy NumberConsider trapezoidal fuzzy number A = (a, b, c, d) is depicted in Fig.2 The mem-bership function µ of A will be defined as follows.
µ(x) =
0 , x < ax−ab−a , a ≤ x ≤ b1 , b ≤ x ≤ cd−xd−c , c ≤ x ≤ d0 , x ≥ d
Publication 1 112
The trapezoidal fuzzy number A = (a, b, c, d) can be represented with an orderpair of function of r-cut approach i.e.[µ(r), µ(r)] = [a + (b− a)r, d− (d− c)r] ,where r ∈ [0, 1]
Fig.2 Trapezoidal fuzzy number
Definition 2.5 Gaussian Fuzzy NumberThe asymmetric Gaussian fuzzy number A = (α, σl, σr).The membership func-tion µ(x) of A will be defined as follows.
µ(x) =
e− (x−α)2
2σ2l , x ≤ α
e− (x−α)2
2σ2r , x ≥ αwhere ,the modal value (center) denote as α and σl, σr denote left and right handspreads(fuzziness i.e.width) corresponding to the Gaussian Distribution.For sym-metric Gaussian fuzzy number the left and right-hand spreads are equal i.e.σl =σr = σ.So symmetric Gaussian fuzzy number may be written as A = (α, σ, σ)
and corresponding function may be defined as µ(x) = e−β(x−α)2
,∀x ∈ R whereβ = 1
2σ2 . The symmetric Gaussian fuzzy number in parametric form can berepresented as
A = [µ(r), µ(r)] =[α−
√− (loger)
β , α+√− (loger)
β
]where r ∈ [0, 1]
Fig.3 Gaussian fuzzy number
For all the above type of fuzzy numbers the left and right bound of fuzzy numberssatisfy the following requirements
1. µ(r) is a bounded monotonic increasing left continuous function over [0, 1],
2. µ(r) is a bounded monotonic decreasing left continuous function [0, 1],
3. µ(r) ≤ µ(r), 0 ≤ r ≤ 1.
Publication 1 113
Definition 2.6 Fuzzy arithmeticFor any arbitrary two fuzzy numbers u =(u(r), u(r)), v = (v(r), v(r)), 0 ≤ r ≤ 1and arbitrary k ∈ R.we define addition,subtraction,multiplication,scalar multi-plication by k.(see in [21])
u+ v = (u(r) + u(r), v(r) + v(r)),u− v = (u(r)− v(r), u(r)− v(r)),
u · v =(min{u(r)v(r), u(r)v(r), u(r)v(r), u(r)v(r)},max{u(r)v(r), u(r)v(r), u(r)v(r), u(r)v(r)})
ku =
{(ku(r), ku(r)), k ≥ 0(ku(r), ku(r)), k < 0
Definition 2.7 Hukuhara-differenceLet x, y ∈ E.If there exists z ∈ E such that x = y + z,then z is called theHakuhara-difference of fuzzy numbers x and y,and it is denoted by z = x y.The sign stands for Hukuhara-difference,and x y 6= x+ (−1)y.Definition 2.8 Hukuhara-differentiabilityLet f : (a, b)→ E and t0 ∈ (a, b).We say that f is Hukuhara-differential at t0,ifthere exists an element f
′(t0) ∈ E such that for all h > 0 sufficiently small,
∃f(t0 + h) f(t0), f(t0) f(t0 − h) and the limits holds(in the metric D)
limh→0
f(t0 + h) f(t0)
h= limh→0
f(t0) f(t0 − h)
h= f
′(t0).
Definition 2.9 Generalized Hukuhara differentiabilityLet f : (a, b) → E and t0 ∈ (a, b).We say that f is (1)-differential at t0,if thereexists an element f
′(t0) ∈ E such that for all h > 0 sufficiently small,
∃f(t0 + h) f(t0), f(t0) f(t0 − h) and the limits holds(in the metric D)
limh→0
f(t0 + h) f(t0)
h= limh→0
f(t0) f(t0 − h)
h= f
′(t0).
and f is (2)-differentiable if for all h > 0 sufficiently small,∃f(t0) f(t0 +h),∃f(t0 − h) f(t0)and the limits(in the metric D)
limh→0
f(t0) f(t0 + h)
−h = limh→0
f(t0 − h) f(t0)
−h = f′(t0).
If f′(t0) exist in above cases then i.e called Generalized fuzzy derivative of f(t).
Definition 2.10 Strongly generalized differentiabilityLet f : (a, b)→ E and t0 ∈ (a, b).We say that f is strongly generalized differen-tial at t0(Bede-Gal differential)if there exist an element f
′(t0) ∈ E such that
(i)for all h > 0 sufficiently small,∃f(t0 + h) f(t0),∃f(t0) f(t0 − h)and the
Publication 1 114
limits(in the metric D)
limh→0
f(t0 + h) f(t0)
h= limh→0
f(t0) f(t0 − h)
h= f
′(t0).
or(ii)for all h > 0 sufficiently small,∃f(t0) f(t0 + h),∃f(t0 − h) f(t0)and thelimits(in the metric D)
limh→0
f(t0) f(t0 + h)
−h = limh→0
f(t0 − h) f(t0)
−h = f′(t0).
or(iii)for all h > 0 sufficiently small,∃f(t0 + h) f(t0),∃f(t0 − h) f(t0)and thelimits(in the metric D)
limh→0
f(t0 + h) f(t0)
h= limh→0
f(t0 − h) f(t0)
−h = f′(t0).
or(iv)for all h > 0 sufficiently small,∃f(t0) f(t0 + h),∃f(t0) f(t0 − h)and thelimits(in the metric D)
limh→0
f(t0) f(t0 + h)
−h = limh→0
f(t0) f(t0 − h)
h= f
′(t0).
(h and −h at denominators mean 1h and 1
−h ,respectively)
Theorem 2.1 [14].Let f : R→ E be a function and denotef(t) = (f(t, r), f(t, r)),foreach r ∈ [0, 1].Then
1. If f is (i)-differentiable,then f(t, r) and f(t, r) are differentiable functionand
f′(t) = (f
′(t, r), f
′(t, r))
2. If f is (ii)-differentiable,then f(t, r) and f(t, r) are differentiable functionand
f′(t) = (f
′(t, r), f
′(t, r))
Definition 2.11 Piecewise Continuous Functionf(t) is piecewise continuous function in a ≤ t ≤ b if there exist a finite numbersof points t1, t2, ......, tN such that f(t) is continuous on each open subintervala < t < t1,t1 < t2,.......,tN < t < b, and has a finite limit as t approaches eachendpoint from the interior of that subinterval.Definition 2.12 Exponential Orderf(t) is of exponential order as t→∞ if there exist real constants K, c, T 3|f(t)| ≤ e−ct, t ≥ T.
Publication 1 115
Definition 2.13 Fuzzy Laplace TransformFuzzy Laplace Transform is an example of integral transform relation of theformF (s) =
∫ baK(t, s)f(t)dt,
where t is time and K(t, s) is kernel of transform which transform f(t) to F (s)i.e. which transform time domain to frequency domain.The most well knownintegral transform is Laplace transformwhere a = 0 and b =∞K(t, s) = e−st
F (s) = limb→∞∫ b0e−stf(t)dt,
limb→∞∫ b0e−stf(t)dt = (limb→∞
∫ b0e−stf(t)dt, limb→∞
∫ b0e−stf(t)dt)
also by using definition of classical Laplace transform:
l[f(t, r)] = limb→∞∫ b0e−stf(t)dt and
l[f(t, r)] = limb→∞∫ b0e−stf(t)dt
then we followL[f(t)] = (l[f(t, r), l[f(t, r))
Theorem 2.2 [4].Let f′(t) be an integrable fuzzy-valued function,and f(t) is
the primitive of f′(t) on [0,∞).Then
L[f′(t)] = sL[f(t)] f(0)
where f is (i)-differentiableorL[f
′(t)] = (−f(0)) (−sL[f(t)])
where f is (ii)-differentiableTheorem 2.3 [4].Let f(t) and g(t) be continuous fuzzy-valued functions andc1, c2 are constants.suppose that f(t)e−st, g(t)e−st are improper fuzzy Riemann-integrable on [0,∞],thenL[(c1f(t)) + (c2g(t))] = (c1L[f(t)]) + (c2L[g(t)]).Theorem 2.4 Let f
′′(t) be integrable fuzzy-valued function,and f(t),f
′(t) are
primitive of f′(t),f
′′(t) on [0,∞].Then
L[f′′(t)] = s2L[f(t)] sf(0) f ′(0)
where f is (i)-differentiable and f′
is (i)-differentiable orL[f
′′(t)] = s2L[f(t)] sf(0)− f ′(0)
where f is (ii)-differentiable and f′
is (ii)-differentiable orL[f
′′(t)] = (−s2)L[f(t)]− sf(0)− f ′(0)
where f is (i)-differentiable and f′
is (ii)-differentiable orL[f
′′(t)] = (−s2)L[f(t)]− sf(0) f ′(0)
where f is (ii)-differentiable and f′
is (i)-differentiableTheorem 2.5 Let f(t) satisfies the condition of existence theorem of Laplacetransform and L[f(t)] = F (s) thenL[tf(t)] = −F ′(s)Hence if f
′(t) satisfies the condition of existence theorem of Laplace transform
then
Publication 1 116
L[tf′(t)] = − d
dsL[f
′(t)] = − d
ds
{sF (s)− f(0)
}= −sF ′(s)− F (s)
similarly for f′′
L[tf′′(t)] = − d
dsL[f
′′(t)] = − d
ds
{s2F (s)− sf(0)− f ′(0)
}= −s2F ′(s)− 2F (s) + f(0)
3 Examples
Example 3.1 Consider Variable coefficient differential equation ty′′ − y′ = −1
subject to initial condition y(0) = 0Consider Variable coefficient fuzzy differential equation where we are consider-ing initial condition and forcing function both triangular fuzzy number
ty′′ − y′ = −1, t ∈ [0, 1]
y(0) = [r − 1, 1− r]1 = [r, 2− r]
By using fuzzy Laplace transform method,we have:L[ty
′′]− L[y
′] = −L[1]
By using FLT of derivative and Differentiation of FLTl[y(t, r)] = r−1
s + rs2 + A
s3
l[y(t, r)] = 1−rs + 2−r
s2 + As3 ,where A is constant.
Hence solution is as follows:y(t, r) = (r − 1) + rt+Bt2
y(t, r) = (1− r) + (2− r)t+Bt2,where B is constant.The y(t, r) and y(t, r) at r ∈ [0, 1] are presented in Fig.4
y(t) = y(t, 1) = y(t, 1) = t+Bt2
0.00
0.05
0.10
t
0.0
0.5
1.0
r
-1.0
-0.5
0.0
0.5
1.0
yHt,rL
Fig.4 Solution y(t,r) by using Triangular fuzzy number
Publication 1 117
Consider Trapezoidal fuzzy number
ty′′ − y′ = −1, t ∈ [0, 1]
y(0) = [r − 1, 2− r]1 = [r, 3− r]
By using fuzzy Laplace transform method,we have:L[ty
′′]− L[y
′] = −L[1]
By using FLT of derivative and Differentiation of FLTl[y(t, r)] = r−1
s + rs2 + A
s3
l[y(t, r)] = 2−rs + 3−r
s2 + As3 ,where A is constant.
Hence solution is as follows:y(t, r) = (r − 1) + rt+Bt2
y(t, r) = (2− r) + (3− r)t+Bt2,where B is constant.The y(t, r) and y(t, r) at r ∈ [0, 1] are presented in Fig.5
y(t, 1) = t+Bt2 , y(t, 1) = 1 + 2t+Bt2
0.00
0.05
0.10
t
0.0
0.5
1.0
r
-1
0
1
2
yHt,rL
Fig.5 Solution y(t,r) by using Trapezoidal fuzzy number
Consider Gaussian fuzzy number
ty′′ − y′ = −1, t ∈ [0, 1]
y(0) = [−√−(2loger),
√−(2loger)]
1 = [1−√
(−2loger), 1 +√
(−2loger)]
By using fuzzy Laplace transform method,we have:L[ty
′′]− L[y
′] = −L[1]
By using FLT of derivative and Differentiation of FLT
l[y(t, r)] =−√−(2loger)s +
1−√
(−2loger)s2 + A
s3
l[y(t, r)] =
√−(2loger)
s +1+√
(−2loger)s2 + A
s3 ,where A is constant.Hence solution is as follows:y(t, r) = [−
√−(2loger)] + [1−
√(−2loger)]t+Bt2
y(t, r) = [√−(2loger)] + [1 +
√(−2loger)]t+Bt2,where B is constant.
The y(t, r) and y(t, r) at r ∈ [0, 1] are presented in Fig.6
y(t) = y(t, 1) = y(t, 1) = t+Bt2
Publication 1 118
0.00
0.05
0.10
t
0.0
0.5
1.0
r
-5
0
5
10
yHt,rL
Fig.6 Solution y(t,r) by Gaussian fuzzy number
Example 3.2 Consider Variable coefficient differential equation ty′′
+2y′+ty = 0
subject to boundary conditionsy(0) = 1y(π) = 0Consider Variable coefficient fuzzy differential equation where we are consider-ing boundary conditions triangular fuzzy number
ty′′
+ 2y′+ ty = 0,
y(0) = [r, 2− r]y(π) = [r − 1, 1− r]
By using fuzzy Laplace transform method,we have:L[ty
′′] + 2L[y
′] + L[ty] = −L[0]
By using FLT of derivative and Differentiation of FLT l[y(t, r)] = (r)tan−1( 1s )
l[y(t, r)] = (2− r)tan−1( 1s )
Hence solution is as follows:y(t, r) = (r) sintty(t, r) = (2− r) sinttThe y(t, r) and y(t, r) at r ∈ [0, 1] are presented in Fig.7
y(t) = y(t, 1) = y(t, 1) = sintt
0.0
0.5
1.0r
01 2
3t
0.0
0.5
1.0
1.5
2.0
yHt,rL
Fig.7 Solution y(t,r) by using Triangular fuzzy number
Consider Trapezoidal fuzzy number
ty′′
+ 2y′+ ty = 0,
y(0) = [r, 3− r]y(π) = [r − 1, 2− r]
Publication 1 119
By using fuzzy Laplace transform method,we have:L[ty
′′] + 2L[y
′] + L[ty] = −L[0]
By using FLT of derivative and Differentiation of FLTl[y(t, r)] = (r)tan−1( 1
s )
l[y(t, r)] = (3− r)tan−1( 1s )
Hence solution is as follows:y(t, r) = (r) sintty(t, r) = (3− r) sinttThe y(t, r) and y(t, r) at r ∈ [0, 1] are presented in Fig.8
y(t, 1) = sintt , y(t, 1) = 2 sintt
0.0
0.5
1.0r
01 2
3
t
0
1
2
3
yHt,rL
Fig.8 Solution y(t,r) by using Trapezoidal fuzzy number
Consider Gaussian fuzzy number
ty′′
+ 2y′+ ty = 0,
y(0) = [1−√−(2loger), 1 +
√−(2loger)]
y(π) = [−√
(−2loger),√
(−2loger)]
By using fuzzy Laplace transform method,we have:L[ty
′′] + 2L[y
′] + L[ty] = 0
By using FLT of derivative and Differentiation of FLTl[y(t, r)] = [1−
√(−2loger)]tan
−1( 1s )
l[y(t, r)] = [1 +√
(−2loger)]tan−1( 1
s )Hence solution is as follows:y(t, r) = [1−
√(−2loger)]
sintt
y(t, r) = [1 +√
(−2loger)]sintt
y(t) = y(t, 1) = y(t, 1) = sintt
0.0
0.5
1.0r
01 2
3
t
0
5
yHt,rL
Fig.9 Solution y(t,r) by using Gaussian fuzzy number
Publication 1 120
4 Result and Discussion
From,Examples 1,2 we see that the solution of second order FIVP and FBVPare depends on the derivative i.e.(i)-differentiable or (ii)-differentiable.Thus,asin above examples,the solution can be adequately chosen among four cases ofthe strongly generalize differentiability.On the other hand,In this new procedureunicity of the solution is lost because we have four possibilities,but flexibility isgained in fuzzy context.In above Examples,for Triangular and Gaussian fuzzynumbers at r = 1 upper bound and lower bounds are same and that is same asExact solution of given differential equation but for Trapezoidal fuzzy number,itis some interval that contain Exact solution as lower bound or upper bound.
5 Conclusion
The Fuzzy Laplace transform method provided solutions to variable coefficientsecond order FIVPs and FBVPs by using the strongly generalize differentiabilityconcept.Here we solved FIVPs and FBVPs by using different types of fuzzy num-bers like Triangular,Trapezoidal and Gaussian.In that Triangular fuzzy numberis easy to use in conclusion where as Trapezoidal fuzzy number required longcalculations and Gaussian fuzzy number include logarithmic function i.e. againdifficult to deal with it if asymmetric Gaussian fuzzy number occurs in calcula-tion.The efficiency of method was described by solving numerical examples.
References
[1] M Barkhordari Ahmadi, Narsis Aftab Kiani, and Nasser Mikaeilvand.Laplace transform formula on fuzzy nth-order derivative and its applicationin fuzzy ordinary differential equations. Soft Computing, 18(12):2461–2469,2014.
[2] O Akın, Tahir Khaniyev, Omer Oruc, and IB Turksen. An algorithm forthe solution of second order fuzzy initial value problems. Expert Systemswith Applications, 40(3):953–957, 2013.
[3] Tofigh Allahviranloo, Saeid Abbasbandy, Soheil Salahshour, andA Hakimzadeh. A new method for solving fuzzy linear differential equa-tions. Computing, 92(2):181–197, 2011.
[4] Tofigh Allahviranloo and M Barkhordari Ahmadi. Fuzzy laplace trans-forms. Soft Computing, 14(3):235–243, 2010.
[5] Tofigh Allahviranloo, E Ahmady, and N Ahmady. Nth-order fuzzy lineardifferential equations. Information sciences, 178(5):1309–1324, 2008.
[6] Tofigh Allahviranloo, E Ahmady, and N Ahmady. A method for solving n thorder fuzzy linear differential equations. International Journal of ComputerMathematics, 86(4):730–742, 2009.
Publication 1 121
[7] Tofigh Allahviranloo, Narsis Aftab Kiani, and M Barkhordari. Towardthe existence and uniqueness of solutions of second-order fuzzy differentialequations. Information sciences, 179(8):1207–1215, 2009.
[8] B Bede and SG Gal. Remark on the new solutions of fuzzy differentialequations. Chaos Solitons Fractals, 2006.
[9] Barnabas Bede and Sorin G Gal. Generalizations of the differentiability offuzzy-number-valued functions with applications to fuzzy differential equa-tions. Fuzzy sets and systems, 151(3):581–599, 2005.
[10] Barnabas Bede, Sorin G Gal, et al. Solutions of fuzzy differential equationsbased on generalized differentiability. Communications in MathematicalAnalysis, 9(2):22–41, 2010.
[11] Barnabas Bede, Imre J Rudas, and Attila L Bencsik. First order linearfuzzy differential equations under generalized differentiability. Informationsciences, 177(7):1648–1662, 2007.
[12] James J Buckley and Thomas Feuring. Fuzzy differential equations. Fuzzysets and Systems, 110(1):43–54, 2000.
[13] James J Buckley and Thomas Feuring. Fuzzy initial value problem for nth-order linear differential equations. Fuzzy Sets and Systems, 121(2):247–255,2001.
[14] Y Chalco-Cano and H Roman-Flores. On new solutions of fuzzy differentialequations. Chaos, Solitons & Fractals, 38(1):112–119, 2008.
[15] Minghao Chen, Congxin Wu, Xiaoping Xue, and Guoqing Liu. On fuzzyboundary value problems. Information Sciences, 178(7):1877–1892, 2008.
[16] Zouhua Ding, Ming Ma, and Abraham Kandel. Existence of the solu-tions of fuzzy differential equations with parameters. Information Sciences,99(3):205–217, 1997.
[17] Didier Dubois and Henri Prade. Towards fuzzy differential calculus part 3:Differentiation. Fuzzy sets and systems, 8(3):225–233, 1982.
[18] Amin Esfahani, Omid Solaymani Fard, and Tayebeh Aliabdoli Bidgoli. Onthe existence and uniqueness of solutions to fuzzy boundary value problems.Ann. Fuzzy Math. Inform, 7(1):15–29, 2014.
[19] Omid Solaymani Fard and Nima Ghal-Eh. Numerical solutions for lin-ear system of first-order fuzzy differential equations with fuzzy constantcoefficients. Information Sciences, 181(20):4765–4779, 2011.
[20] A.Khastan F.Bahrami, K.Ivaz. New result on multiple solutions for nth-order fuzzy differential equations under generalized differentiability. Bound-ary Value Problems, 395714(1):1–13, 2009.
Publication 1 122
[21] Menahem Friedman, Ma Ming, and Abraham Kandel. Numerical solu-tions of fuzzy differential and integral equations. Fuzzy sets and Systems,106(1):35–48, 1999.
[22] NA Gasilov, Afet Golayoglu Fatullayev, SE Amrahov, and A Khastan. Anew approach to fuzzy initial value problem. Soft Computing, 18(2):217–225, 2014.
[23] Nizami Gasilov, Sahin Emrah Amrahov, and Afet Golayoglu Fatullayev.Linear differential equations with fuzzy boundary values. In Application ofInformation and Communication Technologies (AICT), 2011 5th Interna-tional Conference on, pages 1–5. IEEE, 2011.
[24] Nizami Gasilov, Sahin Emrah Amrahov, and Afet Golayoglu Fatullayev.Solution of linear differential equations with fuzzy boundary values. FuzzySets and Systems, 257:169–183, 2014.
[25] DN Georgiou, Juan J Nieto, and Rosana Rodriguez-Lopez. Initial valueproblems for higher-order fuzzy differential equations. Nonlinear Analysis:Theory, Methods & Applications, 63(4):587–600, 2005.
[26] Roy Goetschel and William Voxman. Elementary fuzzy calculus. Fuzzysets and systems, 18(1):31–43, 1986.
[27] Xiaobin Guo, Dequan Shang, and Xiaoquan Lu. Fuzzy approximate so-lutions of second-order fuzzy linear boundary value problems. BoundaryValue Problems, 2013(1):212, 2013.
[28] Mir Sajjad Hashemi and J Malekinagad. Series solution of fuzzy wave-likeequations with variable coefficients. Journal of Intelligent & Fuzzy Systems,25(2):415–428, 2013.
[29] Mir Sajjad Hashemi, J Malekinagad, and Hamid Reza Marasi. Series so-lution of the system of fuzzy differential equations. Advances in FuzzySystems, 2012:16, 2012.
[30] Hossein Jafari, Mohammad Saeidy, and Dumitru Baleanu. The variationaliteration method for solving n-th order fuzzy differential equations. OpenPhysics, 10(1):76–85, 2012.
[31] Osmo Kaleva. Fuzzy differential equations. Fuzzy sets and systems,24(3):301–317, 1987.
[32] Osmo Kaleva. The cauchy problem for fuzzy differential equations. Fuzzysets and systems, 35(3):389–396, 1990.
[33] A Kandel and WJ Byatt. Fuzzy differential equations. PROCEEDINGSOF THE INTERNATIONAL CONFERENCE ON CYBERNETICS ANDSOCIETY, 1978.
Publication 1 123
[34] Abraham Kandel. Fuzzy dynamical systems and the nature of their solu-tions. In Fuzzy Sets, pages 93–121. Springer, 1980.
[35] A Khastan, F Bahrami, and K Ivaz. New results on multiple solutionsfor th-order fuzzy differential equations under generalized differentiability.Boundary Value Problems, 2009(1):395714, 2009.
[36] A Khastan and Juan J Nieto. A boundary value problem for second or-der fuzzy differential equations. Nonlinear Analysis: Theory, Methods &Applications, 72(9):3583–3593, 2010.
[37] Alireza Khastan, Juan J Nieto, and Rosana Rodrıguez-Lopez. Variation ofconstant formula for first order fuzzy differential equations. Fuzzy Sets andSystems, 177(1):20–33, 2011.
[38] V Lakshmikantham and Juan J Nieto. Differential equations in metricspaces: an introduction and an application to fuzzy differential equations.DYNAMICS OF CONTINUOUS DISCRETE AND IMPULSIVE SYS-TEMS SERIES A, 10:991–1000, 2003.
[39] Vangipuram Lakshmikantham and Ram N Mohapatra. Theory of fuzzydifferential equations and inclusions. CRC Press, 2004.
[40] V Laksmikantham, S Leela, and AS Vatsala. Interconnection between setand fuzzy differential equations. Nonlinear Analysis: Theory, Methods &Applications, 54(2):351–360, 2003.
[41] S Siah Mansouri and N Ahmady. A numerical method for solving nth-orderfuzzy differential equation by using characterization theorem. Communi-cation in Numerical Analysis, 2012:12, 2012.
[42] Sankar Prasad Mondal, Sanhita Banerjee, and Tapan Kumar Roy. Firstorder linear homogeneous ordinary differential equation in fuzzy environ-ment. Int. J. Pure Appl. Sci. Technol, 14(1):16–26, 2013.
[43] N.B.Desai. Fuzzy modeling and its applications. M.phil Dessertation, 1997.
[44] Donal O’Regan, V Lakshmikantham, and Juan J Nieto. Initial and bound-ary value problems for fuzzy differential equations. Nonlinear Analysis:Theory, Methods & Applications, 54(3):405–415, 2003.
[45] Madan L Puri and Dan A Ralescu. Differentials of fuzzy functions. Journalof Mathematical Analysis and Applications, 91(2):552–558, 1983.
[46] Komal R.Patel and N.B.Desai. Solution of fuzzy initial value problems.ADIT Journal of Engineering, 12(1):53–57, 2015.
[47] Komal R.Patel and N.B.Desai. Solution of fuzzy initial value problems byfuzzy laplace transform. Internation Conference on Research and Innova-tion in Science Engineering and Technology, 2016.
Publication 1 124
[48] S Salahshour. Nth-order fuzzy differential equations under generalized dif-ferentiability. Journal of Fuzzy Set Valued Analysis, 2011, 2011.
[49] Soheil Salahshour and Tofigh Allahviranloo. Applications of fuzzy laplacetransforms. Soft computing, 17(1):145–158, 2013.
[50] Soheil Salahshour and Elnaz Haghi. Solving fuzzy heat equation by fuzzylaplace transforms. In International Conference on Information Processingand Management of Uncertainty in Knowledge-Based Systems, pages 512–521. Springer, 2010.
[51] Seppo Seikkala. On the fuzzy initial value problem. Fuzzy sets and systems,24(3):319–330, 1987.
[52] Shiji Song and Congxin Wu. Existence and uniqueness of solutions tocauchy problem of fuzzy differential equations. Fuzzy sets and Systems,110(1):55–67, 2000.
[53] Smita Tapaswini and S Chakraverty. Numerical solution of n-th order fuzzylinear differential equations by homotopy perturbation method. Interna-tional Journal of Computer Applications, 64(6), 2013.
[54] Hsien-Chung Wu. The fuzzy riemann integral and its numerical integration.Fuzzy Sets and Systems, 110(1):1–25, 2000.
[55] Zhang Yue and Wang Guangyuan. Time domain methods for the solutionsof n-order fuzzy differential equations. Fuzzy Sets and Systems, 94(1):77–92,1998.
[56] Lotfi A Zadeh. The concept of a linguistic variable and its application toapproximate reasoning—i. Information sciences, 8(3):199–249, 1975.
Publication 1 125
30
PRAJÑĀ - Journal of Pure and Applied Sciences Vol. 24 - 25 : 30 - 37 (2017) ISSN 0975 - 2595
SOLUTION OF NTH ORDER FUZZY INITIAL VALUE PROBLEM
Komal R.Patel1, Narendrasinh B.Desai2*
1Assistant Professor, Applied Sciences and Humanities Department, ITM Universe, Vadodara-390510, Gujarat, India
2Associate Professor, Head of Applied Sciences and Humanities Department, ADIT, V.V.Nagar-388121, Gujarat, India
[email protected], [email protected] *Corresponding author
ABSTRACT
In this paper we consider higher order linear differential equations with fuzzy initial values that occurs in almost all engineering branches. Here ,We find solution for constant coefficient and variable coefficient third order linear differential equation by using method based on properties of linear trasformation.We show that fuzzy problem has unique solution if corresponding crisp Problem has unique solution. We will also prove that if the initial values are triangular fuzzy numbers, then the values of the solution at a given time are also triangular fuzzy numbers. We are going to propose a method to find fuzzy solution. We present three examples, One is homogeneous and another is non-homogeneous linear differential equation with constant coefficient and third is non-homogeneous linear differential equation with variable coefficient (Cauchy-Euler equation) to illustrate applicability of proposed method. We also plot graphs to show difference between exact and fuzzy solutions. This shows that our method is practical and applicable to solve nth order fuzzy initial value problems. AMS Subject Classification code: 34A07:Fuzzy differential equations. Keywords: Fuzzy initial value problem, Fuzzy set, Fuzzy number and Linear transformation.
INTRODUCTION
Fuzzy initial value problem occurs in almost all Engineering branches. The term "Fuzzy differential equation" was put forward for the first time by Kandel and Byatt [1],The fuzzy initial value problem was studied by Seikkala [26].The fuzzy initial problem has been investigated by many authors so far Buckley and Feuring [7,8]; Buckley, Feuring, and Hayashi [9]; Lakshmikantham and Nieto [18]; Bede and Gal [5]; Bede, Rudas, and Bencsik [3]; Perfilieva et al.[25]; Chalco-Cano and Román-Flores[10,11]; Khastan, Bahrami, and Ivaz [14]; Gasilov, Amrahov, and Fatullayev [17]; Khastan,Nieto,Rodríguez-López [16]; Patel, Desai [22].Gasilo et al.[13],Gomes and Barros [4,28] proposed concepts of fuzzy calculus, analogically to classical calculus, and studied fuzzy differential equations in terms of this calculus. Under certain conditions, they established the existence of a solution for the first order fuzzy initial value problem and suggested a solution method. Gasilov et al. [17] benefitted from properties of linear transformations and proposed a method to find fuzzy bunch of Solution functions for linear equation. The method is applicable to higher order linear differential equations with constant coefficients. Most of the researchers assume that derivative in the differential equation as a derivative of a fuzzy function in some sense. In earlier researches the derivative was considered as
Hukuhara derivative. A study in this direction was made by Kaleva [19,20,21]. When Hukuhara derivative is used, then uncertainty of the solution may increase infinitely with time. Furthermore, Bede and Gal [5] showed that a simple fuzzy function, generated by multiplication of differentiable crisp function and a fuzzy number, may not have Hukuhara derivative. In order to overcome this difficulty Bede and Gal [5] developed the generalized derivative concept and after that the studies about this subject were accelerated (Bede[2]); Bede et al [6]; Khastan and Nieto[15]; Chalco-Cano,Román-Flores[10,11,29]. But in case of generalized Hukuhara derivatives there are four different cases for second order fuzzy differential equation. Khastan A.,Gasilov.N.A., Fatullayev.A.G, Amrahov.S.E.[27], found solution of constant coefficient FBVP. Patel, Desai [23,24] solve fuzzy initial value problems by fuzzy Laplace transform. In this paper we consider the fuzzy initial value problem as a set of crisp problem using properties of linear transformations. We solve higher order Cauchy-Euler’s equation. For clarity we explain the
proposed method for third order fuzzy linear differential equations, but the results are true for higher-order equations too. The fuzzy solution proposed by our method coincides with extension principle’s results.
Publication 2 126
31
PRELIMINARIES Definition: Membership function:
A fuzzy set 𝐴 can be defined as a pair of the universal set U and the membership function 𝜇: 𝑈 −> [0,1] for each x ∈ 𝑈, the number 𝜇𝐴 is called the membership degree of x in 𝐴 .
Definition: 𝜶 − 𝒄𝒖𝒕 𝒔𝒆𝒕 :
𝐹𝑜𝑟 𝑒𝑎𝑐ℎ 𝛼 ∈ 0,1 the crisp set 𝐴𝛼 = 𝑥 ∈ 𝑈ǀ𝜇𝐴 𝑥 ≥ 𝛼 is called 𝛼 − 𝑐𝑢𝑡 𝑠𝑒𝑡 of 𝐴 .We use the notation 𝑢 = 𝑢𝐿, 𝛼 , 𝑢𝑅(𝛼) 0 ≤ 𝛼 ≤ 1 to indicate a fuzzy
number in parametric form. We denote 𝑢 = 𝑢𝐿 0 and 𝑢 = 𝑢𝑅 0 to indicate the left and the right end-points of 𝑢 respectively. an α-cut of 𝑢 is an interval [𝑢𝐿, 𝛼 , 𝑢𝑅(𝛼)], which we denote as 𝑢𝛼= [u𝛼, 𝑢 𝛼].
Definition:Fuzzy Number A fuzzy number is a fuzzy set like 𝑈: 𝑅 → 𝐼 = [0,1] which satisfies:
(a) 𝑢 is upper semi-continuous on 𝑅 ∀ ∈ > 0 ∃ 𝛿 > 0 ∋ 𝑢 𝑥 − 𝑢 𝑥0 < ∈, 𝑥 − 𝑥0 < 𝛿
(b) 𝑢 is fuzzy convex i.e.(𝜆𝑥 + 1 − 𝜆 𝑦) ≥min 𝑢 𝑥 , 𝑢 𝑦 ∀𝑥, 𝑦 ∈ 𝑅, 𝜆 ∈ [0,1]
(c) 𝑢 is normal i.e ∃𝑥0 ∈ 𝑅 for which 𝑢(𝑥0) = 1
(d)𝐹𝑜𝑟 𝑒𝑎𝑐ℎ 𝛼 ∈ 0,1 , 𝑠𝑢𝑝𝑝 𝑢 = 𝑥 ∈ 𝑅ǀ𝑢𝛼(𝑥) >0 is support of 𝑢,and its 𝑐𝑙(𝑠𝑢𝑝𝑝 𝑢) is compact.
Definition: Triangular Fuzzy Number:
The Triangular fuzzy number as 𝑢 = (a, b, c) for which 𝑢𝐿, 𝛼 = a + 𝛼 (b − a), 𝑢𝑅(𝛼) = b - 𝛼 (c − b)
and u = a, 𝑢 = c. In geometric interpretations, we refer to the point 𝑏 as a vertex. Let us consider a triangular fuzzy number 𝑢 = (p, 0, q) the vertex of which is 0 (Note that p < 0 and q > 0 in this case).Then 𝑢𝐿(𝛼) = (1−α) p and
𝑢𝑅(𝛼) = (1−α) q and consequently, α-cuts are intervals
[(1 − 𝛼) 𝑝, (1 − 𝛼) 𝑞] = (1 − 𝛼)[𝑝, 𝑞] From the last representation one can see that an α-cut is similar to the interval [p, q] (i.e. to the 0-cut) with similarity coefficient (1−α). We often express a fuzzy number 𝑢 as 𝑢 = 𝑢𝑐𝑟 + 𝑢 𝑢𝑛 (crisp part + uncertainty). Here 𝑢𝑐𝑟 is a number with membership degree 1 and represents the crisp part (the vertex) of 𝑢 ,while 𝑢 𝑢𝑛 represents the uncertain part with vertex at the origin.
For a triangular fuzzy number 𝑢 = (a, b, c) we have 𝑢𝑐𝑟 = 𝑏 and 𝑢 𝑢𝑛 = (𝑎 − 𝑏, 0 , 𝑐 − 𝑏). Properties of Fuzzy Valued Number
For arbitrary 𝑢 =(𝑢 𝑟 , 𝑢 (𝑟)) 𝑣 = (𝑣 𝑟 , 𝑣 (𝑟)),
0 ≤ 𝑟 ≤ 1 and arbitrary 𝑘 ∈ 𝑅. We define addition, subtraction, multiplication, scalar multiplication by 𝑘.
𝑢 + 𝑣 = (𝑢 𝑟 + 𝑣 𝑟 , 𝑢 𝑟 + 𝑣 (r)) 𝑢 − 𝑣 = (𝑢 𝑟 − 𝑣 𝑟 , 𝑢 𝑟 − 𝑣(r))
𝑢 ∙ 𝑣 = (min u r v r , 𝑢 𝑟 𝑣 𝑟 , 𝑢 𝑟 𝑣 r , 𝑢 𝑟 𝑣 𝑟 ,
max u r v r , 𝑢 𝑟 𝑣 𝑟 , 𝑢 𝑟 𝑣 r , 𝑢 𝑟 𝑣 𝑟 )
𝑘𝑢 = (𝑘𝑢 𝑟 , 𝑘𝑢 𝑟 ) 𝑖𝑓 𝑘 ≥ 0
(𝑘𝑢 𝑟 , 𝑘𝑢 𝑟 ) 𝑖𝑓 𝑘 < 0
Definition:Hukuhara difference
Let 𝑥, 𝑦 ∈ 𝐸.If there exists 𝑧 ∈ 𝐸 such that 𝑥 = 𝑦 + 𝑧, then 𝑧 is called the Hukuhara difference of fuzzy numbers x and y, and it is denoted by 𝑧 = 𝑥 Ɵ 𝑦. The Ɵ sign stands for Hukuhara-difference and 𝑥 Ɵ 𝑦 ≠ 𝑥 + −1 𝑦.
Definition:Hukuhara differential
Let 𝑓: (𝑎, 𝑏) → 𝐸 and 𝑡0 ∈ (𝑎, 𝑏) if there exists an element 𝑓 ′(𝑡0) ∈ E such that for all ℎ > 0 sufficiently small,exists 𝑓 𝑡0 + ℎ Ɵ 𝑓(𝑡0), 𝑓 𝑡0 Ɵ𝑓(𝑡0 − ℎ) and the limits holds(in the metric D)
limℎ→0f 𝑡0+h Ɵ f(𝑡0)
ℎ = limℎ→0
f 𝑡0 Ɵ f(𝑡0−h)
ℎ =𝑓 ′(𝑡0)
Here derivative is considered as Hukuhara
derivative.
FUZZY INITIAL VALUE PROBLEMS (FIVPS) In this section, We have described a fuzzy initial value problem (FIVP) and concept of solution. We investigate a fuzzy Initial value problem with crisp linear differential equation and fuzzy initial values. FIVP can arise in modeling of a process the dynamics of which is crisp but there are uncertainties in initial values.
Publication 2 127
32
Consider the nth order fuzzy initial value problem where 𝑏𝑛 𝑥 ≠ 0.
𝑏𝑛 𝑥 𝑦𝑛 + 𝑏𝑛−1 𝑥 𝑦𝑛−1 + ⋯ + 𝑏1 𝑥 𝑦′+𝑏0 𝑥 𝑦 = 𝑓 𝑥
𝑦 𝑙 = 𝐴1
𝑦′ 𝑙 = 𝐴2
𝑦′′ 𝑙 = 𝐴3
……
𝑦𝑛−1 𝑙 = 𝐴𝑛
1
Where 𝐴1 , 𝐴2
, 𝐴3 ,…..,𝐴𝑛
are fuzzy numbers and 𝑏0(𝑥), 𝑏1(𝑥), … , 𝑏𝑛 (𝑥) and 𝑓 (𝑥) are continuous crisp functions or constants and 𝑙 is any integer.Let us represent the initial values as 𝐴1
= 𝑎1 + 𝑎1 , 𝐴2 = 𝑎2+ 𝑎2 ,….., 𝐴𝑛
= 𝑎𝑛+ 𝑎𝑛 where 𝑎1 , 𝑎2 , … . . , 𝑎𝑛 are crisp numbers while 𝑎1 𝑎2 , … … , 𝑎𝑛 are fuzzy numbers. We split the FIVP in Eq.(1) to the following problems: Associated crisp problem (which is non-homogeneous)
𝑏𝑛 𝑥 𝑦𝑛 + 𝑏𝑛−1 𝑥 𝑦𝑛−1 + ⋯ + 𝑏1 𝑥 𝑦′+𝑏0 𝑥 𝑦 = 𝑓 𝑥
𝑦 𝑙 = 𝑎1
𝑦′ 𝑙 = 𝑎2
𝑦′′ 𝑙 = 𝑎3
……
𝑦𝑛−1 𝑙 = 𝑎𝑛
(2)
Homogeneous problem with fuzzy initial values
𝑏𝑛 𝑥 𝑦𝑛 + 𝑏𝑛−1 𝑥 𝑦𝑛−1 + ⋯ + 𝑏1 𝑥 𝑦 ′+𝑏0 𝑥 𝑦 = 0
𝑦 𝑙 = 𝑎1
𝑦′ 𝑙 = 𝑎2
𝑦′′ 𝑙 = 𝑎3 ……
𝑦𝑛−1 𝑙 = 𝑎𝑛
(3)
It is easy to see if 𝑦𝑐𝑟 𝑥 and 𝑦 un(x) are solutions of Eq.(2) and Eq.(3) respectively then 𝑦 (𝑥) = 𝑦𝑐𝑟 𝑥 + 𝑦 un(𝑥) is a solution of the given problem in Eq.(1). Hence, Eq.(1) is reduced to solving a non-homogeneous equation with crisp conditions in Eq.(2) and homogeneous equation with fuzzy initial conditions in Eq.(3). Therefore, we will investigate how to solve Eq.(1).To determine 𝑦 (𝑥) we consider Linear transformation 𝑇: 𝑅𝑛 → 𝑅 , 𝑇 𝑢 =𝑣 𝑢, where 𝑣 is fixed 𝑛 𝑥 𝑛 determinant and 𝑢 = [𝑎1 , 𝑎2 , … 𝑎𝑛 ]𝑇.
THE SOLUTION ALGORITHM The solution algorithm consists of four steps: Represent the initial values as 𝐴1
=𝑎1 + 𝑎1 ,
𝐴2 = 𝑎2 + 𝑎2 ,….. , 𝐴𝑛
= 𝑎𝑛+ 𝑎𝑛 Find linear independent solutions 𝑦1 𝑥 ,
𝑦2 𝑥 ,……,𝑦𝑛 𝑥 of the crisp differential equation 𝑏𝑛 (𝑥)𝑦𝑛 + 𝑏𝑛−1 𝑥 𝑦𝑛−1 +
⋯ 𝑏1 𝑥 𝑦 ′+𝑏0 𝑥 𝑦 = 𝑓(𝑥).Constitute the vector-function 𝑠 𝑥 = (𝑦1 𝑥 , 𝑦2 𝑥 , … . , 𝑦𝑛 (𝑥)), the determinant W and calculate the vector-function
𝑡 𝑥 =𝑠 𝑥 𝑊−1 = 𝑡1 𝑥 , 𝑡2 𝑥 , … . , 𝑡𝑛 𝑥
by formula at 𝑥 = 𝑙. The Wronskain
𝑊 = 𝑦1(𝑙) 𝑦2 𝑙 … 𝑦𝑛 𝑙
⋮ ⋮ ⋱ ⋮𝑦1
𝑛−1(𝑙) 𝑦2𝑛−1 𝑙 … 𝑦𝑛
𝑛−1(𝑙)
and det(𝑊) ≠ 0 ∴ 𝑊 −1 𝑒𝑥𝑖𝑠𝑡.
Find the solution 𝑦𝑐𝑟 𝑥 of the non-homogeneous crisp problem.
The solution of the given problems For homogeneous FIVP
y (x) = 𝑡1(𝑥) 𝑎1 + 𝑡2(𝑥) 𝑎2 + … +(𝑥) 𝑎𝑛 . For non-homogeneous FIVP 𝑦 (x) = 𝑦𝑐𝑟 𝑥 + 𝑡1(𝑥) 𝑎1 + 𝑡2(𝑥) 𝑎2 + … +𝑡𝑛 (𝑥) 𝑎𝑛 . EXAMPLES Example:5.1 Solve the 3rd order FIVP with constant coefficient homogeneous equation.
𝑦 ′′′ + 3𝑦 ′′ + 3𝑦 ′ + 𝑦 = 0
𝑦 0 = (−0.5,0,1)
𝑦 ′ 0 = (−1,0,1)
𝑦 ′′ 0 = (−1,0,0.5)
(4)
The problem is homogeneous and initial values are fuzzy numbers with vertices at 0.therefore solution by solution algorithm.𝑦 𝑥 = 𝑒−𝑥 , 𝑦2 𝑥 =𝑥𝑒−𝑥and 𝑦3 𝑥 = 𝑥2𝑒−𝑥are linearly independent solution for the equation 𝑦 ′′′ + 3𝑦 ′′ + 3𝑦 ′ + 𝑦 = 0. Hence 𝑠(𝑥) = (𝑒−𝑥 , 𝑥𝑒−𝑥 , 𝑥2𝑒−𝑥) and
Publication 2 128
33
𝑊 = 1 0 0
−1 1 01 −2 2
and
𝑡(𝑥) = 𝑠 𝑥 𝑊−1 = (𝑡1(𝑥), 𝑡2(𝑥), 𝑡3(𝑥)),
𝑡(𝑥) = (𝑒−𝑥 + 𝑥𝑒−𝑥 +𝑥2
2𝑒−𝑥 , 𝑥𝑒−𝑥
+ 𝑥2𝑒−𝑥 ,𝑥2
2𝑒−𝑥)
𝑦 𝑢𝑛 𝑥 = 𝑒−𝑥 + 𝑥𝑒−𝑥 + 𝑥2𝑒−𝑥 −0.5,0,1
+ 𝑥𝑒−𝑥 + 𝑥2𝑒−𝑥 (−1,0,1)
+𝑥2
2𝑒−𝑥(-1,0,0.5)
Where the arithmetic operations are considered to be fuzzy operations. The fuzzy solution 𝑦 𝑥 form band in the 𝑥𝑦-coordinate space (Fig.1) Since the initial values are triangular fuzzy numbers, an α-cut of the solution can be determined by similarity coefficient (1−α), i.e. 𝑦𝛼 𝑥 = 1 − 𝛼 [y 𝑥 , 𝑦 𝑥 ]
Fig.1 The fuzzy solution, obtained by the proposed method, for Example 5.1 Example:5.2 Consider the 3rd order FIVP with constant coefficient non-homogeneous equation.
𝑦 ′′′ + 3𝑦 ′′ + 3𝑦 ′ + 𝑦 = 30𝑒−𝑥
𝑦 0 = 2.5,3,4
𝑦 ′ 0 = −4, −3, −2
𝑦 ′′ 0 = −48, −47, −46.5
5
We represent the initial values as 𝐴 = (2.5,3,4) = 3 + (-0.5,0,1), 𝐵 = (-4,-3,-2) = -3 + (-1,0,1), 𝐶 =(-48,-47,-46.5) = - 47 + (-1,0,0.5).
we solve crisp non-homogeneous crisp problem
𝑦 ′′′ + 3𝑦 ′′ + 3𝑦 ′ + 𝑦 = 30𝑒−𝑥
𝑦 0 = 3
𝑦 ′ 0 = −3
𝑦 ′′ 0 = −47
(6)
And the crisp solution 𝑦𝑐𝑟 𝑥 = 3 − 25𝑥2 𝑒−𝑥 + 5𝑥3𝑒−𝑥
𝑦𝑐𝑟 𝑥 = 3 𝑒−𝑥 + 𝑥𝑒−𝑥 + 𝑥2𝑒−𝑥 − 3 𝑥𝑒−𝑥 + 𝑥2𝑒−𝑥
−47𝑥2
2𝑒−𝑥 + 5𝑥3𝑒−𝑥
Fuzzy homogeneous problem to find the uncertainty of
the solution is as follows:
𝑦 ′′′ + 3𝑦 ′′ + 3𝑦 ′ + 𝑦 = 0
𝑦 0 = (0.5,0,1)
𝑦 ′ 0 = (−1,0,1)
𝑦 ′′ 0 = (−1,0,0.5)
(7)
This problem is the same as Example 1. Hence, the solution is
𝑦 𝑢𝑛 𝑥 = 𝑒−𝑥 + 𝑥𝑒−𝑥 + 𝑥2𝑒−𝑥 −0.5,0,1 + 𝑥𝑒−𝑥 + 𝑥2𝑒−𝑥 (−1,0,1)
+𝑥2
2𝑒−𝑥(-1,0,0.5)
Fig. 2 The fuzzy solution 𝑦 𝑥 , obtained by the proposed method, for Example 5.2.Red line represents the crisp solution.
Publication 2 129
34
We add this uncertainty to the crisp solution and get the fuzzy solution of the given FIVP
𝑦 𝑥 = 𝑦𝑐𝑟 𝑥 + 𝑦 𝑢𝑛 𝑥
= 𝑒−𝑥 + 𝑥𝑒−𝑥 + 𝑥2𝑒−𝑥 2.5,3,4 + 𝑥𝑒−𝑥 + 𝑥2𝑒−𝑥 (−4, −3, −2)
+𝑥2
2𝑒−𝑥(-48,-47,-46.5) + 5𝑥3𝑒−𝑥
In above example if we take 𝑥 → ∞ then fuzziness in solution is disappeared. So the solution goes nearer to zero if we increase 𝑥.(see Fig.3)
Fig. 3 If x increases fuzziness disappears Example:5.3 Consider 3rd order FIVP with variable coefficient.
(𝑥3𝐷3 − 3𝑥2𝐷2 + 6𝑥𝐷 − 6)𝑦 =
12
𝑥
𝑦 1 = 4,5,6
𝑦′ 1 = 12,13,14.5
𝑦′′ 1 = 9.5,10,11
(8)
(𝑥3𝐷3 − 3𝑥2𝐷2 + 6𝑥𝐷 − 6)𝑦 =
12
𝑥
𝑦 1 = 5
𝑦 ′ 1 = 13
𝑦 ′′ 1 = 10
(9)
Associated non-homogeneous problem has the solution i.e. exact solution of non-homogeneous problem
𝑦𝑐𝑟 𝑥 =−138
24𝑥 +
328
24𝑥2 +
−69
24𝑥3 −
1
24𝑥
Consider homogeneous equation
(𝑥3𝐷3 − 3𝑥2𝐷2 + 6𝑥𝐷 − 6)𝑦 = 0
𝑦 1 = −1,0,1
𝑦 ′ 1 = −1,0,0.5
𝑦 ′′ 1 = −0.5,0,1
(10)
The problem is homogeneous and initial values are fuzzy numbers with vertices at 0.therefore solution by solution algorithm.𝑦1 𝑥 = 𝑥 , 𝑦2 𝑥 = 𝑥2 and 𝑦2 𝑥 = 𝑥3 are linearly independent solution of equation ( 𝑥3𝐷3 − 3𝑥2𝐷2 + 6𝑥𝐷 − 6)𝑦 = 0 Hence 𝑠 𝑥 =(𝑥, 𝑥2 , 𝑥3 ) and
𝑊 = 1 1 11 2 30 2 6
And 𝑡(𝑥)= 𝑠 𝑥 𝑊−1 = (𝑡1(𝑥), 𝑡2(𝑥), 𝑡3(𝑥))
𝑡 𝑥 = 3𝑥 − 3𝑥2 + 𝑥3 , −2𝑥 + 3𝑥2 − 𝑥3 ,𝑥
2− 𝑥2 +
𝑥3
2
Fig.4 The fuzzy solution, obtained by the proposed method, for Example 5.3 The fuzzy solution is
𝑦 𝑢𝑛 𝑥 = 3𝑥 − 3𝑥2 + 𝑥3 −1,0,1 + −2𝑥 + 3𝑥2 − 𝑥3 −1,0,0.5
+ (𝑥
2− 𝑥2 +
𝑥3
2)(−0.5,0,1)
We add this uncertainty to the crisp solution and get the fuzzy solution of the given FIVP
𝑦 𝑥 = 𝑦𝑐𝑟 𝑥 + 𝑦 𝑢𝑛 𝑥 = 3𝑥 − 3𝑥2 + 𝑥3 4,5,6
+ −2𝑥 + 3𝑥2 − 𝑥3 (12, 13, 14.5)
+( 𝑥
2− 𝑥2 +
𝑥3
2)(9.5,10,11) -
1
24𝑥
Publication 2 130
35
𝑦 𝑥 = −6𝑥 + 14𝑥2 − 3𝑥3 −1
24𝑥
Fig.5 The fuzzy solution, obtained by the proposed method, for Example 5.3.Red line represents the crisp solution. In above example if we increase 𝑥 and as we take 𝑥 → ∞ then fuzziness in solution increases and goes to infinite.
Fig. 6 If we increase 𝑥 fuzziness increases Example:5.4 Consider a unit mass sliding on a frictionless table attached to a spring, with spring constant k =16. Suppose the mass is lightly tapped by a hammer every T seconds. Suppose that the first tap occurs at time t = 0 and before that time the mass is at rest. Describe what happens to the motion of the mass for the tapping period T =1.
Fig. 7
𝑥 ′′ 𝑡 + 16 𝑥 𝑡 = 𝑠𝑖𝑛𝑡𝑥 0 = 0
𝑥 ′ 0 = 1
(11)
Consider the homogeneous equation with fuzzy parameters.
𝑥 ′′ 𝑡 + 16 𝑥 𝑡 = 0 𝑥 0 = (−1,0,1)
𝑥 ′ 0 = (−0.5,0,1)
(12)
The problem is homogeneous and initial values are fuzzy numbers with vertex 0 therefore solution by solution algorithm. 𝑥1(𝑡) = 𝑐𝑜𝑠4𝑡 , 𝑥2(𝑡) = 𝑠𝑖𝑛4𝑡 are linearly independent solution for the equation 𝑥 ′′(𝑡) + 16𝑥(𝑡) = 0 Hence 𝑠(𝑡) = (𝑐𝑜𝑠4𝑡, 𝑠𝑖𝑛4𝑡) and
𝑊 = 1 00 4
and 𝑦(𝑡) = (𝑦1(𝑡), 𝑦2(𝑡)) Here 𝑦(𝑡) = (𝑐𝑜𝑠4𝑡, (1/ 4) 𝑠𝑖𝑛4𝑡) 𝑥 𝑢𝑛 𝑡 = (cos4t)(−1,0,1) + (1/ 4)sin4t(−0.5,0,1) Crisp solution of given non-homogeneous problem 𝑥𝑐𝑟 (𝑡) = (7/ 30)𝑠𝑖𝑛4𝑡 + (1/15)𝑠𝑖𝑛𝑡 . 𝑥 (𝑡) = 𝑥𝑐𝑟 (𝑡)+𝑥 𝑢𝑛 𝑡
=𝑐𝑜𝑠4𝑡(−1,0,1)+(7/30)𝑠𝑖𝑛4𝑡(0.5,1,2) + (1/15)𝑠𝑖𝑛𝑡 = (7/ 30)𝑠𝑖𝑛4𝑡 + (1/ 15)𝑠𝑖𝑛𝑡 where arithmetic operations are considered to be fuzzy operations. the fuzzy solution 𝑥 (𝑡)form band in the tx-coordinate space.(Fig.8)
Fig.8:The periodic fuzzy solution by properties of Linear transformation
Publication 2 131
36
If we apply periodic force to block, it will move in given open interval that we denote by triangular fuzzy number, Here the damping force is negligible. The spring mass system occurs in all most all engineering branches as well as in the many real life situations. CONCLUSIONS In this paper we have represented the fuzzy initial value problems as a set of crisp problems. We have proposed a solution method based on the properties of linear transformations. For clarity we have explained the proposed method for third order linear differential equation but it is applicable for nth order also. Here we have solved third order Cauchy-Euler equation i.e. we solved variable coefficient initial value problems based on the properties of linear transformations. We also solved one application level problem. AIM AND SCOPE OF THE WORK The aim of paper is to find solution of any Fuzzy differential equation having uncertain initial conditions. In future we are going to solve Fuzzy differential equations having the applications in different branches of engineering. we will solve higher order FBVP based on the properties of linear transformations. REFERENCES Periodical articles
[1] Kandel A., Byatt W. J. , ―Fuzzy differential equations,.‖ in Proceedings of the International Conference on Cybernetics and Society, (1978.): pp. 1213–1216, Tokyo, Japan, .
[2] B, Bede. ―A note on ―two-point boundary value problems associated with non-linear fuzzy differential equations‖.‖
Fuzzy Sets Syst (2006): 157:986–989.
[3] Bade, B, I,J. Rudas and A.L. Bencsil. ―First order linear
fuzzy differential equations under generalized differentiability.‖ Information Sciences, (2007):177:1648-1662.
[4] Barros L C, Gomes L T, Tonelli P A. ― Fuzzy differential
equations: an approach via fuzzification of the derivative operator.‖ Fuzzy Sets Syst. (doi:10.1016/j.fss.2013.03.004).
[5] Bede B, Gal S G. ―Generalizations of the differentiability of
fuzzy number valued functions with applications to fuzzy differentia lequation.‖ Fuzzy Sets Syst (2005): 151:581–
599.
[6] Bede B, Stefanini L. ―Generalized differentiability of fuzzy-valued functions.‖ Fuzzy Sets Syst. (2012): doi:10.1016/j.fss.2012.10.003.
[7] Buckley JJ, Feuring T. ― Fuzzy differential equations.‖
Fuzzy Sets Syst (2000): 110:43–54.
[8] —. ―Fuzzy initial value problem for Nth-order linear differential equation.‖ Fuzzy Sets Syst (2001): 121:247–
255.
[9] Buckley JJ, Feuring T, Hayashi Y. ― Linear systems of first
order ordinary differentia lequations:fuzzy initial conditions.‖ Soft Comput (2002): 6:415–421.
[10] Chalco-Cano Y, Román-Flores H. ― On the new solution of
fuzzy differential equations.‖ Chaos Solitons Fractals (2008): 38:112–119.
[11] —. ―Comparation between some approaches to solve fuzzy
differential equations.‖ Fuzzy Sets Syst (2009): 160(11):1517–1527 .
[12] Desai, N. B. ―Fuzzy modeling and its application.‖ M.Phil Dessertation (1997).
[13] Gasilov N A, Amrahov ¸SE, Fatullayev AG. ― A geometric approach to solve fuzzy linear systems of differential equations.‖ Appl Math Inf Sci (2011a): 5:484–495 .
[14] Khastan A, Bahrami F, Ivaz K. ―New results on multiple
solutions for Nth-order fuzzy differential equations under generalized differentiability.‖ Bound Value Probl. (2009): doi:10.1155/2009/395714.
[15] Khastan A, Nieto JJ. ―A boundary value problem for second
order fuzzy differential equations.‖ Nonlinear Anal (2010): 72:3583–3593 .
[16] Khastan A, Nieto JJ, Rodríguez-López R. ―Variation of
constant formula for first order fuzzy differential equations. ‖ Fuzzy Sets Syst (2011): 177:20–33.
[17] Gasilov N. A., Fatullayev A. G. ¸ Amrahov S. E., Khastan A ―A new approach to fuzzy initial value problem.‖ Soft Comput (2013): DOI 10.1007/s00500-013-1081-z.
[18] Lakshmikantham, V. and J.J. Nieto. ―Differential equations
in metric spaces:an introduction and an application to fuzzy differential equations.‖ Dynamics of (2003): A, vol. 10, no. 6, pp. 991–1000.
[19] Kaleva O., ―A note on fuzzy differentia lequations.‖ Non linear Anal (2006): 64:895–900 .
[20] —. ―Fuzzy differential equations.‖ Fuzzy Sets Syst. (1987): 24:301– 317 .
[21] —. ―The Cauchy problem for fuzzy differential equations. .‖
Fuzzy Sets Syst (1990): 35:389–396.
[22] Patel K. R, Desai N.B. ―Solution of fuzzy initial value
problems.‖ ADIT Journal of Engineering (December 2015): Vol.12,No.1,53-57.
[23] Patel K. R., Desai N. B. ―Solution of variable coefficient
fuzzy differential equations by fuzzy Laplace transform.‖
IJRITCC Vol.5,Issue:6 (June 2017): 927-942.
Publication 2 132
37
[24] Patel K.R, Desai N.B. ―Solution Of Fuzzy Initial Value
Problems By Fuzzy Laplace Transform.‖ Kalpa Publications in Computing,ICRISET2017 (Vol.2,June 12, 2017): 25-37.
[25] Perfilieva I, Meyer H,Baets B,Plšková D.
―Cauchyproblemwith fuzzy initial condition and its approximate solution with the help of fuzzy transform. .‖ In: IEEE world congress on computational intelligence (WCCI 2008), (2008): p 6.
[26] Seikkala, S. ―On the fuzzy initial value problem.‖ Fuzzy Sets and Systems, (1987): vol. 24, no. 3, pp. 319–330.
Conference proceedings
[27] Gasilov N, Amrahov ¸SE, Fatullayev AG. ―Linear
differential equations with fuzzy boundary values.‖ In: Proceedings of the 5th international conference―Application
of Information and Communication Technologies‖ (2011b): (AICT2011) (ISBN: 978-1-61284-830-3), pp 696–
700.Baku, Azerbaijan, October 12–14.
[28] Gomes L T, BarrosL C. ―Fuzzy calculus via extension of
the derivative and integral operators and fuzzy differential equations.‖ In:Proceedings of annual meeting of the NAFIPS (North American Fuzzy Information Processing Society) (2012): pp 1–5 .
[29] Chalco-Cano Y, Román-Flores H,Rojas-Medar MA. ―Fuzzy
differential equations with generalized derivative.‖
In:Proceedings of 27th NAFIPS international conference IEEE (n.d.)
Publication 2 133