60.1HJHUL /DEDQJ 0DGXUD ,QGRQHVLD
Transcript of 60.1HJHUL /DEDQJ 0DGXUD ,QGRQHVLD
Journal of Physics: Conference Series
PAPER • OPEN ACCESS
Subject Scheduling Using Genetic Algorithms (Case Study: SMK Negeri1 Labang-Madura-Indonesia)To cite this article: Eka Mala Sari Rochman et al 2020 J. Phys.: Conf. Ser. 1569 022074
View the article online for updates and enhancements.
This content was downloaded from IP address 120.188.38.79 on 25/07/2020 at 03:16
Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distributionof this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
Published under licence by IOP Publishing Ltd
International Conference on Science and Technology 2019
Journal of Physics: Conference Series 1569 (2020) 022074
IOP Publishing
doi:10.1088/1742-6596/1569/2/022074
1
Subject Scheduling Using Genetic Algorithms (Case Study:
SMK Negeri 1 Labang-Madura-Indonesia)
Eka Mala Sari Rochman1, Muhammad Ali Syakur2, Imamah3, Aeri Rachmad4
Faculty of Engineering, University of Trunojoyo Madura. Raya Telang Road P.O. BOX 2 Kamal, Bangkalan-Madura,Indonesia
[email protected], [email protected]
Abstract. The manually scheduling system is considered less effective and efficient because it
requires a long time. Problems will become more complex if the number of components or data
used is increasing. The schedule is expected not only not to experience clashes, but also to adjust
to some limitations that must be met. Genetic Algorithm is one of the heuristic search algorithms
that are very well used in solving optimization problems. The problem of scheduling genetic
algorithms is considered to have good performance in finding the optimal solution. Genetic
Algorithms implement an evolutionary process by randomly producing chromosomes from each
population These chromosomes produce a solution to the problem raised, namely scheduling
subjects. The conclusion of this study is to be able to arrange the schedule of subjects efficiently,
by overcoming obstacles such as clashing schedules without eliminating the constraints that must
be met. Keywords: Schedulilng System, optimization problems Genetic Algorithm.
1. Introduction
The definition of a schedule is the division of time based on a work order arrangement plan; list or table
of activities or plan of activities with a detailed division of the time of implementation, while the
definition of scheduling is a process, method, action schedule or include in the schedule [1]. So
scheduling is something that must be present in every activity. Scheduling techniques vary depending
on the size, complexity, duration of the activity [2] [3]. One activity that requires scheduling is an
activity in the world of education in the form of a schedule of subjects. The lesson schedule is used as a
guideline by teachers, students or principals in carrying out the teaching and learning process.
Understanding of the lesson schedule is a set of learning sequences that are used as guidelines and must
be followed in the process of teaching and learning activities. Labang 1 State Vocational School is a vocational high school that has a problem in terms of unique
scheduling which has 27 classes, there are 6 vocational majors with 4 types of subjects namely
normative, adaptive, mulok (Local Content) and vocational subjects. In the process of teaching and
learning activities SMK Negeri 1 Labang has a total of 50 lesson hours per week for each class with a
total of teachers. There are several aspects that must be considered to get an effective and efficient lesson
schedule. The related aspects of scheduling include: no occurrence of clash of subjects in the same class
and at the same time, no teacher who teaches subjects more than one different class at the same time
and the number of hours of subjects in accordance with total time of week.
The method for making lesson schedules that has been done so far is still manual, that is by doing it
by searching for any columns that are still empty, then placing a schedule in that column. The schedule
International Conference on Science and Technology 2019
Journal of Physics: Conference Series 1569 (2020) 022074
IOP Publishing
doi:10.1088/1742-6596/1569/2/022074
2
produced in this way takes a long time and tends to ignore these various aspects. Therefore it is necessary
to develop a lesson scheduling system that can accommodate various aspects that are considered above.
There are several methods and algorithms that can be used to solve scheduling problems where each
algorithm has advantages and disadvantages. One method and algorithm is Genetic Algorithm. Genetic
Algorithm is one algorithm that adopts the mechanism of natural genetics that was developed to find
solutions to problems such as scheduling [4] [5]. Genetic algorithms find solutions, namely in the
initialization process. Processing is done to get a solution improvement [6]. With the use of genetic
algorithms, scheduling will be obtained in accordance with the appropriate aspects.
2. Methodology
2.1. Data Collection
The data used in this study include subject data, teacher data, class data, study hours and data
assignments of teaching teachers in the 2017-2018 school year at 1 Labang Vocational High School.
Table 1 shows the input data used.
Table 1 Data
No Data Input Jumlah
1 Teacher 44 Teachers
2 Subjects 20 Subjects
3 Class 9 classes
4 Time allocation 50 hours / week
2.2. Genetic Algorithm
Genetic Algorithm is an optimization method to find the optimal solution to a problem. Genetic
algorithms are search algorithms that are based on natural system mechanisms namely genetic and
natural selection [7] [8] [5]. In contrast to conventional search techniques, genetic algorithms depart
from a set of randomly generated solutions.
Figure 1. Flowchart of genetic algorithms for scheduling
This set is called a population. Whereas every individual in a population is called a chromosome which
is a representation of a solution. Chromosomes evolve in a continuous iterative process called
generation. In each generation, chromosomes are evaluated based on an evaluation function. After
several generations the genetic algorithm will converge on the best chromosome, which is expected to
be the optimal solution. Genetic Algorithms have a simple optimization methodology as follows:
1. Determine the population of a certain number of solutions
Start
Schedule Data
GA Process:
1. Chromosome Representation
2. Generating the initial population
3. Selection
4. Crossover
5. Mutation
End
Schedule of subjects
International Conference on Science and Technology 2019
Journal of Physics: Conference Series 1569 (2020) 022074
IOP Publishing
doi:10.1088/1742-6596/1569/2/022074
3
2. Calculating fitness value functions all solutions in the population
3. Choose several solutions with the highest function fitness value
4. Perform optimization by means of mutations and crossover as much as needed
5. Determine the best solution as a solution to the problem being optimized
Finish Steps:
1) Initial Parameters
The parameters specified in the genetic algorithm are:
a. Fitness function (objective function) that is owned by each candidate scheduling to determine the
individual's fitness level with the criteria to be achieved.
b. Population number of individuals involved in each generation through a selection process.
c. The maximum number of generations that occurs
d. The probability of a cross (pc) in a generation.
e. The probability of a mutation (pm) in each individual.
2) Encoding Techniques
The coding technique is a method of coding / initializing genes from chromosomes. Each
chromosome contains a number of genes containing information stored on the chromosome. In this
study using coding techniques in the form of string bit / varchar which is used in genetic
programming.
3) Set Population and Chromosome Representations
Determining the initial population is the process of generating a number of chromosomes randomly.
Chromosomes represent one possible alternative solution. Chromosomes can be said to be the same
as individuals. Population size depends on the problem to be resolved. After population size is
determined, then initial population generation is carried out by initializing possible solutions into a
number of chromosomes. The length of one chromosome is determined based on the problem under
study.
The length of one chromosome is the multiplication of the number of lesson hours per week with
the number of classes available. One gene contains information on subjects, classes, and lesson
hours. For example for initialization of chromosome formation, for example there is a distribution
of subjects in table 2, distribution of time in table 3.
4) Calculation of Fitness Value
Individuals in the population have been formed, the next step is to calculate the fitness value of each
individual. Fitness will be calculated based on the number of violations found in the individual. The
value generated from the function indicates how optimal the solution is. In the case of subject
scheduling, the smaller the number of violations produced, the better the solution will be.
F = 1 / 1 + (∑ BG + ∑ BK + ∑ BJ) (1) Keterangan :
BG : Teacher clashes at the same time
BK : Clash of classes at the same time
International Conference on Science and Technology 2019
Journal of Physics: Conference Series 1569 (2020) 022074
IOP Publishing
doi:10.1088/1742-6596/1569/2/022074
4
BJ : Clash of lesson hours
Limits used to calculate fitness values:
1. Teachers may not teach more than once at the same time
2. One class may not be scheduled more than once at the same time
3. One subject may not be scheduled more than once at the same time
5) Selection
In this study the method used is roulette wheel selection. In this selection, individuals are selected based
on their fitness values to choose which individuals will experience the marriage process or crossovers,
the better the quality of an individual the greater the chance to be chosen. In the roulette wheel selection
process, the cumulative value of the fitness probability of each individual will be calculated. Examples
of selection steps using roulette wheel:
a. Calculate the total fitness value
b. Calculate the probability of each chromosome, how: fitness value / total fitness
c. Look for intervals for each chromosome
d. Generate 0-1 random numbers as many as existing chromosomes Eg 0.75; 0.25; 0.9; 0.5
e. Search for each random number at which chromosome interval.
f. Then the results of the new population obtained from the selection are
Then the results of the new population obtained from the selection are
Chromosome 1 Chromosome 3
Chromosome 2 Chromosome 1
Chromosome 3 Chromosome 4
Chromosome 4 Chromosome 2
g. Use the selected population for the next process
6) Crossover
Crossover is done by crossing 2 mothers to produce offspring. In the crossing process, the method used
is one-cut-point crossover by taking several genes from the first parent which are then combined with
the genes in the second parent. The child number of the reproductive process is determined by
determining the pc value (crossover probability) which is random from a value between 0 and 1.
Examples of crossover steps:
a. Determine pc value
b. Generate random numbers for each chromosome
c. Search for random numbers that are less than a pc
d. Awaken random numbers (1 to the length of the chromosome) to determine the cut point.
e. Perform crossover, for example the crossover that will be done is chromosome 2 and 4, with cut point
2
Chromosome 1 Chromosome 2
PAI6 TKJ2 T1 MUL4 TKJ2 T21
BIG4 TKJ1 T9 PJS3 TKJ2 T18
MUL3 TKJ2 T5 FIS5 TKJ2 T15
FIS3 TKJ1 T9 BIG1 TKJ1 T15
BIG6 TKJ2 T15 P1D4 TKJ1 T15
BIG3 TKJ1 T19 FIS2 TKJ1 T3
P1D2 TKJ1 T17 PAI6 TKJ2 T23
FIS1 TKJ1 T19 SIM1 TKJ1 T15
f. Use the crossover population for the next process
International Conference on Science and Technology 2019
Journal of Physics: Conference Series 1569 (2020) 022074
IOP Publishing
doi:10.1088/1742-6596/1569/2/022074
5
7) Mutations
Mutation is a process of reproduction by modifying the arrangement of genes from a parent. From one
mutation process to one parent will produce one child. Before the reproduction process is carried out
using mutations, it is necessary to determine the value of pm (probability of mutation), which is
generated from a value between 0 and 1, which is the function of pm to determine how many children
must be produced in the mutation process. The steps of the mutation process:
a. Calculate total genes, total genes = many genes on one x chromosome many chromosomes in the
population
b. Many genes are transferred = total gene x pm. For example, the total genes = 16 and pm = 0.1, many
genes that are mutated are 1.6 ≈ 2
c. Generate random numbers 0 to 1 as many genes as possible
d. Search for genes with random numbers < pm
e. If the results are less than mutated genes, then all are used. If more than the maximum mutated gene,
random to take as many genes that are mutated.
f. Perform the mutation process, by changing the time component in the gene
Before mutation BIG4 TKJ1 T9
After mutation BIG4 TKJ1 T3
g. Use the mutation population for the next process
8) Stop Conditions
Iteration will continue to run and will stop if one of the following conditions has been fulfilled
a. After several successive generations the best fitness value of the population has not changed
b. The maximum number of generations has been reached
c. If the minimum best fitness value has been reached. If there is no violation of the chromosome
3. Result and Discussion
In this test the author uses several parameters that will be applied to the genetic algorithm process. Test
parameters can be seen in the following table 4 below:
Table 4 Testing Parameters
Parameter Value
Number of Individuals 10
Crossover probability 0.7
Mutation Probability 0.1
Many Generation 10
Figure 2. Data Figure 3. Generate Schedule
International Conference on Science and Technology 2019
Journal of Physics: Conference Series 1569 (2020) 022074
IOP Publishing
doi:10.1088/1742-6596/1569/2/022074
6
Figure 4. Generate Results
CONCLUSIONS
Based on the results and discussion, the conclusions that can be obtained from this study are as follows:
1. The performance of genetic algorithms is good enough for subject scheduling
2. To achieve optimization it is best influenced by several parameters applied
3. The more number of generations that are raised, the fewer the number of crashes that occur, the better
the solution will be.
Table 5 Accuracy
Generation Total Crash Accuracy
10 32 93.03 %
20 31 93.13 %
50 23 93.17 %
80 19 94.00 %
100 16 95.88 %
ACKNOWLEDGMENT The authors state that there is no issue of interest with the agencies listed in this paper. We would like
to thank our colleagues at the University of Trunojoyo Madura - Indonesia, the Laboratory of Network
and Multimedia team who helped author in completing this research. It gave the author opportunities
in using the equipment of the laboratory
References
[1] P. D. R. Z. Mudra S. Gondane, "Staff Scheduling in Health Care Systems," Journal of
Mechanical and Civil Engineering (IOSRJMCE) , vol. 1, no. 6, pp. 28-40, 2012.
[2] S. M. H. A.-H. M. e.-L. Y. A.-M. Hesham Abdel-Khalek, "Financing - Scheduling Optimization
for Construction Projects by using Genetic Algorithms," International Journal of Economics and
Management Engineering, pp. 1337-1345, Vol:5, No:11, 2011 .
[3] M. A. M. Imam Ahmad Ashari, "Comparison Performance of Genetic Algorithm and Ant Colony
Optimization in Course Scheduling Optimizing," Scientific Journal of Informatics , vol. 3, no. 2,
pp. 149-158, 2016.
[4] W. M. a. D. B.-s. Firas Albalas, "Optimized Job Scheduling Approach based on Genetic
Algorithms in Smart Grid Environment," International Journal of Communication Networks and
Information Security (IJCNIS) , vol. Vol. 9, no. No.2, p. 172, August 2017 .
International Conference on Science and Technology 2019
Journal of Physics: Conference Series 1569 (2020) 022074
IOP Publishing
doi:10.1088/1742-6596/1569/2/022074
7
[5] S. D. KUMAR RITWIK, "A GENETIC ALGORITHM-BASED APPROACH FOR
OPTIMIZATION OF SCHEDULING IN JOB SHOP ENVIRONMENT," Journal of Advanced
Manufacturing Systems , vol. 10, no. 2, p. 223–240 , 2011.
[6] M. D. Kidwell and D. J. Cook, "Genetic Algorithm for Dynamic Task Scheduling," IEEE, pp. 61-
67, 1994.
[7] W. F. M. Vivi Nur Wijayaningrum, "Optimization of Ship’s Route Scheduling Using Genetic
Algorithm," Indonesian Journal of Electrical Engineering and Computer Science , vol. 2, no. 1,
pp. 180-1861, 2016.
[8] S. M. A. K. Vijaya Bhosale, "A Review of Genetic Algorithm used for optimizing scheduling of
Resource Constraint construction projects," International Research Journal of Engineering and
Technology (IRJET), vol. 4, no. 5, pp. 2869-2872, 2017.
[9] "Optimization of Ship’s Route Scheduling Using Genetic Algorithm," Indonesian Journal of
Electrical Engineering and Computer Science , vol. 2, no. 1, pp. 180-186, 2016.
ICST 2019 International Conference on Science and Technology (ICST) is a scientific forum to discuss the development and latest research on exact science, engineering and technology. The following topics of thid conference are Physical Sciences and Chemistry, Mathematics and Statistics, Informatics and Computer Science, Aerospace Engineering, Biomedical Engineering, Chemical Engineering, Civil Engineering, Computer Engineering, Electrical Engineering and Electronics, Power and Energy Engineering, Environmental Engineering, Geological and Geophysical Engineering, Industrial Engineering, Manufacturing and Mechanical Engineering, Engineering Physics and Materials Engineering, Mining and Petroleum Engineering. Technology-based applications such as learning media for education can also be accepted at this Conference. Selected papers that have been through the peer review will be published in IOP and indexed by Scopus. We respectfully request that you join us to ICST 2019. We look forward to welcoming you to Surabaya! In 2019, ICST is focusing on Science, Technology and Engineering related to the 4th industrial revolution known as Industry 4.0. Therefore, the theme of this conference is
Industrial Revolution 4.0 : The opportunities and challenges in elevating local resources contribute to develop a better world All accepted papers will be published by
and indexed by Scopus
1.1. COMMITTEES • Honorary General Chairs: • Prof. Ir.
Renanto Handogo, M.Sc., Ph.D., Institut Teknologi Sepuluh Nopember Surabaya, Indonesia • Prof. dr. ir. G.J. (Bart) Verkerke, Rijksuniversiteit Groningen, The Netherland • Assoc. Professor Hadi Susanto, M.Sc., Ph.D., University of Essex, United Kingdom • Dr. Jamari, S.T., M.T., University of Diponegoro, Indonesia • General Chairs: • Prof. Dr. Arif Muntasa., S.Si., M.T.,
University of Trunojoyo Madura, Indonesia • Advisory Board:
• Prof. Nobuo Funabiki, Ph.D., Okayama University, Japan
• Prof. Dr. Ir. Muhammad Idrus Alhamid, University of Indonesia, Indonesia • Prof. Dr. Madlazim, M.Si, Universitas Negeri Surabaya, Indonesia • Prof. Dr. rer. nat. Ir. A. P. Bayuseno, M.Sc., University of Diponegoro, Indonesia • Prof. Dr. Sari Edi Cahyaningrum, M.Si., Universitas Negeri Surabaya, Indonesia • Prof. Dr. Ir. Rachmad Hidayat., M.T., IPU., AER., University of Trunojoyo Madura, Indonesia • Prof. T. M. Indra Mahlia, Ph.D., University of Technology Sidney, Australia • Prof. Dr. Revolson A. Mege, M.S., State University of Manado, Indonesia • Roman Podraza, Ph.D., Warsaw University of Technology, Poland • Dr. Dani Harmanto, University of Derby, United Kingdom • Dr. Ario Sunar Baskoro, S.T, M.T, M.Eng., University of Indonesia, Indonesia • Dr. Dra. Jariyah, MP., UPN “Veteran” East Java, Indonesia • Dr. Ir. Ni Ketut Sari, M.T., UPN “Veteran” East Java, Indonesia • Dr. Ir. Reda Rizal, M.Si., UPN "Veteran" Jakarta, Indonesia • Dr. Maspiyah, M.Kes., Universitas Negeri Surabaya, Indonesia • Saiful Anwar, STP, MP., State Polytechnic of Jember, Indonesia • Lita Asyriati Latif, S.T., M.TM., Universitas Khairun Ternate, Indonesia • Ir. Made Mudhina, M.T., Bali State Polytechnic, Indonesia
• Daniel Parenden, S.T., M.T., University of Musamus Merauke , Indonesia • Scientific Committee (also reviewer):
• Wahyudi Agustiono, S.Kom., M.Sc., Ph.D, University of Trunojoyo Madura, Indonesia • Dr. Hakam Muzakki, S.T., M.T., University of Trunojoyo Madura, Indonesia • Chandra Ade Irawan, Ph.D., University of Nottingham Ningbo, China • Abdul Malek Bin Yakoob, Ph.D., Universiti Utara Malaysia, Malaysia • I Dewa Made Cipta Santosa, ST, M.Sc, PhD., Bali State Polytechnic, Indonesia • Muhammad Yusuf, S.T., M.MT., Ph.D., University of Trunojoyo Madura, Indonesia • Dr. Basuki Rahmat, S.Si., M.T., UPN "Veteran" East Java, Indonesia • Dr. Iis Hamsir Ayub Wahab, S.T., M.Eng., Universitas Khairun Ternate, Indonesia • Reviewers (Out of Scientific Committee):
• Prof. Dr. Arif Muntasa, S.Si., M.T., University of Trunojoyo Madura, Indonesia • Prof. Dr. Madlazim, M.Si., Universitas Negeri Surabaya, Indonesia • Prof. Dr. Revolson A. Mege, M.S., State University of Manado, Indonesia • Dr. Dani Harmanto, University of Derby, United Kingdom • Dr. Ario Sunar Baskoro, S.T., M.T., M.Eng., University of Indonesia, Indonesia • Dr. Indah Agustien S., S.Kom. M.Kom., University of Trunojoyo Madura, Indonesia • Dr. Dra. Jariyah, MP., UPN “Veteran” East Java, Indonesia • Dr. Ir. Ni Ketut Sari, M.T., UPN “Veteran” East Java, Indonesia • Dr. Ir. Reda Rizal, M.Si., UPN "Veteran" Jakarta, Indonesia • Dr. Maspiyah, M.Kes., Universitas Negeri Surabaya, Indonesia • Dr. Akhmad Sabaruddin, S.T., M.T., University of Trunojoyo Madura, Indonesia • Dr. Muhammad Amin, S.T., M.T., University of Samudra, Indonesia • Dr. Iis Hamsyir Ayub Wahab, S.T., M.Eng., Universitas Khairun Ternate, Indonesia • Dr. Mufti Amir Sultan, M.T., Universitas Khairun Ternate, Indonesia • Dr. Damora Rhakasywi, S.T., M.T., UPN "Veteran" Jakarta, Indonesia • Dr. Ir. Halim Mahfud, M.Sc., UPN Veteran Jakarta, Indonesia • Dr. Dira Ernawati, S.T., M.T., UPN "Veteran" East Java, Indonesia • Dr. I Gede Susrama, M.Kom., UPN "Veteran" East Java, Indonesia • Dr. Philipus Betaubun, M.T., Universitas Musamus Merauke, Indonesia • Dr. Ir. Budi Hariono, M.Si., State Polytechnic of Jember, Indonesia • Dr. Lilik Anifah, S.T., M.T., Universitas Negeri Surabaya, Indonesia • Dr. Yeni Kustiyahningsih, S.Kom., M.Kom., University of Trunojoyo Madura, Indonesia • Dr. Noor Ifada, S.T., MISD., University of Trunojoyo Madura, Indonesia • Dr. Edy Winarno, S.T., M.Eng., Unisbank, Semarang, Indonesia • Dr. Ir. Lilik Sudiajeng, M.Erg., Bali State Polytechnic, Indonesia • Dr. Anak Agung Ngurah Gde Sapteka, S.T., M.T., Bali State Polytechnic, Indonesia • Dr. Fuad Hasan, S.P., M.P., University of Trunojoyo Madura, Indonesia • Dr. Erina Rahmadyanti, S.T., M.T., Universitas Negeri Surabaya, Indonesia • Rusli Yosfiah, Ph.D., Universiti Teknologi Malaysia, Malaysia • Euis Nurul Hidayah, S.T., M.T., Ph.D., UPN Veteran East Java, Indonesia • Ida Bagus Irawan Purnama, S.T., M.Sc., Ph.D., Bali State Polytechnic, Indonesia • I Gst.Ngr.A. Dwijaya Saputra, S.T., M.T., Ph.D., Bali State Polytechnic, Indonesia • Putu Wijaya Sunu, S.T., M.T., Bali State Polytechnic, Indonesia • Mahrus Khoirul Umami, S.T., M.Sc., Ph.D., University of Trunojoyo Madura, Indonesia • Eppy Yundra, S.Pd., M.T., Ph.D., Universitas Negeri Surabaya, Indonesia • Cahyo Indarto, S.TP., M.P., Ph.D., University of Trunojoyo Madura, Indonesia • Dr. Kukuh Winarso, S.Si, M.T., University of Trunojoyo Madura, Indonesia