02 or Optimization-01 Chapter-04

download 02 or Optimization-01 Chapter-04

of 6

Transcript of 02 or Optimization-01 Chapter-04

  • 8/10/2019 02 or Optimization-01 Chapter-04

    1/6

    Dr. Vishnu Prasad PandeyMaster Degree Program, Purbanchal University

    2014.08.07

    LECTURE #02

    Introduction to Optimization (1)

    MEM124 Operations Research

    Lecture#02: Operational Research

    Contents of Chapter 4

    4. Introduction to Optimization [9 hrs]

    Linear and Multi-objective optimization models;

    Modeling optimization problem in EXCEL; Building linear programing models;

    Solving linear programing models;

    Interpreting solver results and sensitivity analysis;

    Solving multi-objective models;

    Using premium solver for linear programming

    Lecture#02: Operational Research

    Contents of this lecture

    4.Introduction to Optimization

    Introduction

    Optimization modeling process

    Solution of optimization models

    Building linear programing models;

    Solving linear programing models;

    Graphical method

    Simplex method

    Lecture#02: Operational Research

    Optimization: Introduction

    A mathematical Optimization model consists of

    An Objective Function (OF), &

    A set of constraints in a form of system of equations

    inequalities

    It seeks to answer the question What is the best?

    Answer can be expressed as numerical value

    Optimization models are used extensively in almost all

    areas of decision-making, e.g. Engineering design

    Physical, chemical & biological sciences

    Economics

    Management, financial portfolio selection

  • 8/10/2019 02 or Optimization-01 Chapter-04

    2/6

    Lecture#02: Operational Research

    Optimization: Modeling Procedure

    . Describe theproblem

    . Prescribe a

    solution, &

    . Control the

    problem byassessing/updating the

    optimal solutioncontinuously,while changing

    the parameters &structure of the

    problem

    Problemformulation

    Lecture#02: Operational Research

    Optimization Model (OM): Basic contents

    OMs are prescriptive or normative models seeking to findthe best possible strategy for decision-makers

    Optimization problems are made up of following basicingredients:

    Decision variables: the controllable inputs which affe

    the value of the objective function

    An objective funct ion: that we want to mini/maximize

    Parameters: uncontrollable inputs which may be fixed

    numbers

    Constraints: relations between decision variables & thparameters

    Lecture#02: Operational Research

    Optimization: Solutions

    Feasible solution

    A solution value for decision variables, where all the

    constraints are satisfied

    Most solution algorithms proceed by

    first finding a feasible solution,

    then seeking to improve upon it, and

    finally changing the decision variables to move from one

    feasible solution to another feasible solutionThe process is repeated until the OF has reached its

    maximum or minimum

    Optimal solution:

    Gives the minimum or maximum result of the OF while

    satisfying the constraints

    Lecture#02: Operational Research

    Optimization: Solutions

    Optimization problems are classified & solved according t

    mathematical characteristics of the OF,

    constraints, &

    controlled decision variables

    Linear Programming (LP) deals with: a case of optimizatio

    problem, where both the OF & the constraints are lineariterms of the decision variables

    Multi-objective Program (MP) or Goal Program, is: where a single objective characteristic of an optimizati

    problem is replaced by several goals

    in solving MP, one may represent some of the goals as

    constraintsto be satisfied, while other objectives can b

    weightedto make a composite single OF

  • 8/10/2019 02 or Optimization-01 Chapter-04

    3/6

    Lecture#02: Operational Research

    Linear Programming (LP): Building LP

    LP is a mathematical procedure for determining optimalallocation of scarce resources

    LP deals with a class of programming problems where boththe OF & constraints are linear & all relations among the

    variables corresponding to resources are linear

    While building a LP model:

    What are decision variables?

    What are parameters?

    What is the objective function?

    What are the constraints?

    Lecture#02: Operational Research

    Example#01 (same as in Lecture#1)

    The Two Mines Company own two different mines that produceore which, after being crushed, is graded into three classes: higmedium and low-grade. The company has contracted to provide

    smelting plant with 12 tons of high-grade, 8 tons of medium-graand 24 tons of low-grade ore per week. The two mines have

    different operating characteristics as detailed below

    Mine Cost/day ('000) Production (tons/day)

    High Medium Low

    A 180 6 3 4

    B 160 1 1 6

    If maximum operating hours/week of the mines will be 5, formulate

    LP model to answer How many days per week should each mibe operated to fulfill the smelting plant contract?

    Lecture#02: Operational Research

    Example#01: Formulation of LP

    Decision variables:

    x = days/week mine A is operated

    y = days/week mine B is operated

    Objective function (to minimize the cost):

    min Z = 180x + 160y

    Constraints:

    High: 6x + 1y >= 12

    Medium: 3x + 1y >= 8

    Low: 4x + 6y >= 24

    x

  • 8/10/2019 02 or Optimization-01 Chapter-04

    4/6

    Lecture#02: Operational Research

    Solving LP: Extreme point solution method

    Coordinates of all corners (or extreme) points of feasible regionare determined and values of OF at these points are computed &compared

    Because: optimal solution to any LP problem always lie at one of

    the corner points of feasible solution space

    Steps/Procedure:

    Plot each constraint on a graph paper, by considering all & are

    =; and then plot the line

    For each line, divide the region into 3 parts. Pick a point in either

    side of the line & plug its coordinates into the constraint. If it

    satisfies the condition, this side is feasible; otherwise the other side

    is feasible. For equality constraints, only the points on the line are

    feasible.

    Throw away the sides that are NOT feasible

    After all the constraints are graphed, you should have a non-empty

    (convex) feasible region, unless the problem is infeasible

    Lecture#02: Operational Research

    Solving LP: Example#01

    Find graphical solution for a LP model formulated as below:

    OF: Maximize profit function (7T + 5C) [T=Table, C=Chair]

    S.T. Constraints: 3T + 4C 100; T, C > 0

    SOLUTION:

    Plot graphs for eachconstraints

    Constraint line 1

    (carpentry): 3T + 4C =

    2400

    Intercepts (for T = 0, C =600; for T = 800, C = 0) 0 T

    600

    C

    0

    Feasible

    < 2400 hrs

    Infeasible

    > 2400 hrs

    Lecture#02: Operational Research

    Solving LP: Example#01

    SOLUTION:

    Constraint line 2(painting): 2T + 1C =

    1000

    Intercepts:

    for T = 0, C = 1000;

    for T = 500, C = 0

    0 500 800 T

    C

    1000

    600

    0

    Lecture#02: Operational Research

    Solving LP: Example#01

    SOLUTION:

    Maximum cost for chair:C = 450.

    Minimum cost for Table:

    T = 100.

    C

    1000

    600

    450

    0

  • 8/10/2019 02 or Optimization-01 Chapter-04

    5/6

    Lecture#02: Operational Research

    Solving LP: Example#01

    OF line: 7T + 5C =

    Profit.

    Lecture#02: Operational Research

    Solving LP: Example#01

    If additionalconstraints were

    available, such as,

    need at least 75more chairs than

    tables, then: C > T+ 75; or C T >

    75.

    New constraint

    line would have

    been added asshown in the

    graph & Optimal

    point would have

    been shifted to anew point

    T = 320

    C = 360

    No longe

    feasible

    New optimal point

    T = 300, C = 375

    Lecture#02: Operational Research

    Solving LP: Special cases of LP

    Redundant Constraints do not affect the feasible region

    Example: If two constraints are given as X 10 & X

    12, the second constraint is redundant as it is less

    restrictive

    Infeasibility when no feasible solutions exist (there is nofeasible region)

    Example: if X 10 & X 15

    Lecture#02: Operational Research

    Solving LP: Special cases of LP

    Al ternate optimalsolution : when there are

    more than one optimal

    solutions

    Example:

    Maximize (2T + 2C)

    Subject to:

    T + C < 10

    T < 5

    C < 6

    T, C > 0

    0 5 10

    C

    10

    6

    0

    All points onRedsegment are

    optimal

  • 8/10/2019 02 or Optimization-01 Chapter-04

    6/6

    Lecture#02: Operational Research

    Solving LP: Special cases of LP

    Unbounded solutions:when nothing prevents

    the solution from

    becoming infinitely large

    Example:

    Maximize (2T + 2C)

    Subject to:

    2T + 3C > 6

    T, C > 0

    C

    2

    1

    0

    Lecture#02: Operational Research

    Class Activities: Graphical methods

    1. Solve the following LP problem using Graphical method:

    Objective function: Max Z = 6x1 - 4x2

    Subject to: 2x1+ 4x24;

    4x1+ 8x216;

    x1, x20

    2. Solve the following LP problem using Graphical method:

    Objective function: Max Z = 15x1 + 10x2

    Subject to:

    4x1+ 6x2360;

    3x1180;

    5x2200 & x1, x20

    Lecture#02: Operational Research

    Thank you!

    Thank You!!(For enquiries: [email protected])