Sie sind auf Seite 1von 17

Operations Research

Session 2
Solving the Merton Truck problem
Dr. Jyotirmoy Dalal

OR model

OR
terminologies

Maximize Z = 3000 x1 +
5000 x2

Objective Function

Subject to
x1 + 2 x2 <= 4000
2 x1 + 2 x2 <= 6000
2 x1

Constraints

<= 5000
3 x2 <=

4500
x1, x2 >= 0

Decision Variables

Few more terminologies


Feasible region
Solution
Feasible solution
Infeasible solution
Optimal solution
Optimal objective value

x2

Graphical Solution Method


(in 2-D Space)

x1 >=0

(0,0)

x2 >= 0

x1

x2

Adding constraint for Engine Assembly

Which half space


to consider?
2000

+ 2 x2 <=
x1
4000
Or

Half space 2

+ <= 1

Half space 1
4000
(0,0)
What is the feasible region after adding this constraint?

x1

x2
3000

Adding constraint for Metal Stamping

2000
Half space 1

2x1
+ 2 x2 <=
6000

Half space 2

Or
+ <= 1

(?,?)

4000
(0,0)

3000

What is the feasible region after adding this constraint?

x1

x2
(0,3000)

Adding constraint for M 101 assembly


2x1 <=
5000

(0,2000)

Or
X1 <=
2500
(2000,1000)
(?,?)

[do we need to calculate it?]

(?,?)
(4000,0)
(0,0)

(2500,0)(3000,0)

x1

x2

Adding constraint for M 102 assembly

(0,3000)

3x1 <=
4500
(0,2000)

(?,?)

Or
X1 <=
1500

(0,1500)
(2000,1000)
Feasible Region
For the problem

(2500,500)

(4000,0)
(0,0)

(2500,0)(3000,0)

x1

x2
(0,3000)

Complete Feasible region for Merton Truck Problem

Z = 3000 x1 + 5000 x2
(0,2000)

(1000,1500)

How to draw it?

(0,1500)
(2000,1000)
Feasible Region
For the problem

(2500,500)

(4000,0)
(0,0)

(2500,0)(3000,0)

x1

Zoom in near (0,0) point


Z = 3000 x1 + 5000 x2
Let Z = 0, x1 = 1, then x2 = -3/5

(0,0)

(1,0
)

(0,-0.6)
This is Z =0 line

x2

In search of optimal solution

Verify:
Which Constraints are binding?

(1000,1500)

(0,1500)
(2000,1000)
Feasible Region
For the problem

(0,0)

(2500,500)

(2500,0)

x1

x2

An alternate approach
(complete enumeration of all corner

Corner Points
Z =?
------------------------------------------------------(0,0)
points)
(2500,0)

(1000,1500)

(2500,500)
(2000,1000)

(0,1500)
(1000,1500)

(2000,1000)

(0,1500)
Feasible Region
For the problem

(0,0)

(2500,500)

(2500,0)

Is it enough to evaluate the CORNER POINTS only to find optimal solution?

x1

x2

Special case 1: Alternate Optima

Modified Objective Function 1

Modified Objective Function 2

Feasible Region
For the problem

Original Objective function

x1

x2

Special case 2: Unbounded Solution Space

Is this an unbounded space?

x1

x2

Special case 2: Unbounded Solution Space

Solution : bounded or unbounded?

x1

x2

Solution : bounded or unbounded?

x1

x2

Special case 3: Infeasible solution

x1

Das könnte Ihnen auch gefallen