math 527 deterministic or

48
MATH 527 Deterministic OR Graphical Solution Method for Linear Programs

Upload: chelsi

Post on 04-Jan-2016

25 views

Category:

Documents


1 download

DESCRIPTION

MATH 527 Deterministic OR. Graphical Solution Method for Linear Programs. 30. 20. 10. 4. 12. 20. 30. 20. 10. 4. 12. 20. 30. 20. 10. 4. 12. 20. 30. 20. 10. 4. 12. 20. 30. 20. 10. 4. 12. 20. 30. 20. 10. 4. 12. 20. 30. 20. 10. 4. 12. 20. 30. 20. 10. 4. - PowerPoint PPT Presentation

TRANSCRIPT

Page 1: MATH 527  Deterministic OR

MATH 527 Deterministic OR

Graphical Solution Method for Linear Programs

Page 2: MATH 527  Deterministic OR

MATH 327 - Mathematical Modeling

2

4 12 20

10

20

30

6025 21 xx

Page 3: MATH 527  Deterministic OR

MATH 327 - Mathematical Modeling

3

4 12 20

10

20

30

6025 21 xx

Page 4: MATH 527  Deterministic OR

MATH 327 - Mathematical Modeling

4

4 12 20

10

20

30

6025 21 xx

Page 5: MATH 527  Deterministic OR

MATH 327 - Mathematical Modeling

5

4 12 20

10

20

30

6025 21 xx

2021 xx

Page 6: MATH 527  Deterministic OR

MATH 327 - Mathematical Modeling

6

4 12 20

10

20

30

6025 21 xx

2021 xx

Page 7: MATH 527  Deterministic OR

MATH 327 - Mathematical Modeling

7

4 12 20

10

20

30

6025 21 xx

2021 xx

Page 8: MATH 527  Deterministic OR

MATH 327 - Mathematical Modeling

8

4 12 20

10

20

30

6025 21 xx

2021 xx

01 x

Page 9: MATH 527  Deterministic OR

MATH 327 - Mathematical Modeling

9

4 12 20

10

20

30

6025 21 xx

2021 xx

01 x

02 x

Page 10: MATH 527  Deterministic OR

MATH 327 - Mathematical Modeling

10

4 12 20

10

20

30

6025 21 xx

2021 xx

01 x

02 x

Feasible region

The feasible region is a polygon!!

Page 11: MATH 527  Deterministic OR

MATH 327 - Mathematical Modeling

11

How do we find the optimal solution?? We must graph the

isoprofit line.– Straight line– All points on the line

have the same objective value

– When problem is minimization, called an isocost line.

How??– Choose any point in

the feasible region– Find its objective

value (or z-value)– Graph the line

objective function = z-

value.

Page 12: MATH 527  Deterministic OR

MATH 327 - Mathematical Modeling

12

4 12 20

10

20

30

Isoprofit linez = 300

Page 13: MATH 527  Deterministic OR

MATH 327 - Mathematical Modeling

13

4 12 20

10

20

30

Isoprofit line

Page 14: MATH 527  Deterministic OR

MATH 327 - Mathematical Modeling

14

4 12 20

10

20

30

Isoprofit line

Page 15: MATH 527  Deterministic OR

MATH 327 - Mathematical Modeling

15

4 12 20

10

20

30

Isoprofit line

Page 16: MATH 527  Deterministic OR

MATH 327 - Mathematical Modeling

16

4 12 20

10

20

30

Isoprofit line

Page 17: MATH 527  Deterministic OR

MATH 327 - Mathematical Modeling

17

4 12 20

10

20

30

Isoprofit linez = 433 1/3

optimal solution: (20/3, 40/3)z = 433 1/3

Page 18: MATH 527  Deterministic OR

MATH 327 - Mathematical Modeling

18

Binding vs. Nonbinding

A constraint is binding if the optimal solution satisfies that constraint at equality (left-hand side = right-hand side). Otherwise, it is nonbinding.

Binding constraints keep us from finding better solutions!!

Page 19: MATH 527  Deterministic OR

MATH 327 - Mathematical Modeling

19

4 12 20

10

20

30

optimal solution: (20/3, 40/3)z = 433 1/3

Page 20: MATH 527  Deterministic OR

MATH 327 - Mathematical Modeling

20

4 12 20

10

20

30

optimal solution: (20/3, 40/3)z = 433 1/3

binding

Page 21: MATH 527  Deterministic OR

MATH 327 - Mathematical Modeling

21

4 12 20

10

20

30

optimal solution: (20/3, 40/3)z = 433 1/3

binding

binding

Page 22: MATH 527  Deterministic OR

MATH 327 - Mathematical Modeling

22

Convex Sets

A set of points S is a convex set if the line segment joining any two points in S lies entirely in S

ConvexNonconvex

Page 23: MATH 527  Deterministic OR

MATH 327 - Mathematical Modeling

23

Extreme Points

A point P in a convex set S is an extreme point if, for any line segment containing P which lies entirely in S, P is an endpoint of that segment.

A

B

C

D

Page 24: MATH 527  Deterministic OR

MATH 327 - Mathematical Modeling

24

Extreme Points

A point P in a convex set S is an extreme point if, for any line segment containing P which lies entirely in S, P is an endpoint of that segment.

A

B

C

D

C and D are extreme pointsA and B are not

Page 25: MATH 527  Deterministic OR

MATH 327 - Mathematical Modeling

25

Interesting Facts

The extreme points of a polygon are the corner points.

The feasible region for any linear program will be a convex set.

Page 26: MATH 527  Deterministic OR

MATH 327 - Mathematical Modeling

26

Interesting Facts

The feasible region will have a finite number of extreme points

Extreme points are the intersections of constraints (including nonnegativity)

Any linear program that has an optimal solution has an extreme point that is optimal!!

What are the implications?

Page 27: MATH 527  Deterministic OR

MATH 327 - Mathematical Modeling

27

2 6 10

4

8

12

1826 21 xx

Page 28: MATH 527  Deterministic OR

MATH 327 - Mathematical Modeling

28

2 6 10

4

8

12

1826 21 xx

Page 29: MATH 527  Deterministic OR

MATH 327 - Mathematical Modeling

29

2 6 10

4

8

12

1826 21 xx

Page 30: MATH 527  Deterministic OR

MATH 327 - Mathematical Modeling

30

2 6 10

4

8

12

1826 21 xx

3053 21 xx

Page 31: MATH 527  Deterministic OR

MATH 327 - Mathematical Modeling

31

2 6 10

4

8

12

1826 21 xx

3053 21 xx

Page 32: MATH 527  Deterministic OR

MATH 327 - Mathematical Modeling

32

2 6 10

4

8

12

1826 21 xx

3053 21 xx

Page 33: MATH 527  Deterministic OR

MATH 327 - Mathematical Modeling

33

2 6 10

4

8

12

1826 21 xx

3053 21 xx

01 x

Page 34: MATH 527  Deterministic OR

MATH 327 - Mathematical Modeling

34

2 6 10

4

8

12

1826 21 xx

3053 21 xx

01 x02 x

Page 35: MATH 527  Deterministic OR

MATH 327 - Mathematical Modeling

35

2 6 10

4

8

12

1826 21 xx

3053 21 xx

01 x02 x

Feasible Region

Page 36: MATH 527  Deterministic OR

MATH 327 - Mathematical Modeling

36

2 6 10

4

8

12

Isocost linez = 54

Page 37: MATH 527  Deterministic OR

MATH 327 - Mathematical Modeling

37

2 6 10

4

8

12

Isocost line

Page 38: MATH 527  Deterministic OR

MATH 327 - Mathematical Modeling

38

2 6 10

4

8

12

Isocost line

Page 39: MATH 527  Deterministic OR

MATH 327 - Mathematical Modeling

39

2 6 10

4

8

12

Isocost line

Page 40: MATH 527  Deterministic OR

MATH 327 - Mathematical Modeling

40

2 6 10

4

8

12

Isocost linez = 36 1/4

optimal solution: (5/4, 21/4)z = 36 1/4

Page 41: MATH 527  Deterministic OR

MATH 327 - Mathematical Modeling

41

Special Cases

So far, our models have had– One optimal solution– A finite objective value

Does this always happen?

What if it doesn’t?

Page 42: MATH 527  Deterministic OR

MATH 327 - Mathematical Modeling

42

Special Case # 1: Unbounded Linear Programs If maximizing: there are points in the

feasible region with arbitrarily large objective values.

If minimizing: there are points in the feasible region with arbitrarily small objective values.

Page 43: MATH 527  Deterministic OR

MATH 327 - Mathematical Modeling

43

Special Case #1: Unbounded Linear Programs

maximization minimization

Page 44: MATH 527  Deterministic OR

MATH 327 - Mathematical Modeling

44

CAUTION!!!

There is a difference between an unbounded linear program and an unbounded feasible region!!!

Page 45: MATH 527  Deterministic OR

MATH 327 - Mathematical Modeling

45

Special Case #2: Infinite Number of Optimal Solutions When isoprofit/isocost lie intersects an

entire line segment corresponding to a binding constraint

Occurs when isoprofit/isocost line is parallel to one of the binding constraints

Page 46: MATH 527  Deterministic OR

MATH 327 - Mathematical Modeling

46

Special Case #2: Infinite Number of Optimal Solutions

Page 47: MATH 527  Deterministic OR

MATH 327 - Mathematical Modeling

47

Special Case # 3: Infeasible Linear Program Feasible Region is empty

Page 48: MATH 527  Deterministic OR

MATH 327 - Mathematical Modeling

48

Every Linear Program

Has a unique optimal solution, or…..

Has infinite optimal solutions, or…..

Is unbounded, or…..

Is infeasible.