0% found this document useful (0 votes)
10 views

Goal Programming1

The document describes using goal programming to optimize production quantities of two products given resource constraints. There are two objectives - maximize profit and minimize production time to reduce overtime. Multiple solutions are possible as the objectives conflict. Goal programming is used to incorporate both objectives and provide an acceptable solution.

Uploaded by

ms22a031
Copyright
© © All Rights Reserved
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
10 views

Goal Programming1

The document describes using goal programming to optimize production quantities of two products given resource constraints. There are two objectives - maximize profit and minimize production time to reduce overtime. Multiple solutions are possible as the objectives conflict. Goal programming is used to incorporate both objectives and provide an acceptable solution.

Uploaded by

ms22a031
Copyright
© © All Rights Reserved
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 13

Goal Programming

A company makes 2 products which is sold at Rs 800 and Rs 900.


Two resources are required and 200 and 210 units are available per
week. The requirements for the products are 4 and 6 of the first
resource and 5 and 3 of the second resource.

There is a commitment to meet total demand of 40.

Let X1 and X2 be the quantities produced.

Maximize 800X1 + 900X2


4X1 + 5X2 ≤ 200
6X1 + 3X2 ≤ 210
X1 + X2 ≥ 40
X1 , X 2 ≥ 0
6X1 + 3X2 ≤ 210
(0, 40) Z = 36000
(25, 20) Z = 38000
(30, 10) Z = 33000

(0, 40)
Z = 36000

(25, 20)
X1 + X2 ≥ 40

(30, 10) 4X1 + 5X2 ≤ 200


It takes 50 minutes to make the first product and 70 minutes for
product 2. Only 2400 minutes are available in a week and the rest is
to be treated as overtime. We wish to minimize time used (so that
OT can be reduced)
Let X1 and X2 be the quantities produced.

Minimize 50X1 + 70X2 (0, 40) Z = 2800


(25, 20) Z = 2650
4X1 + 5X2 ≤ 200 (30, 10) Z = 2200
6X1 + 3X2 ≤ 210
X1 + X2 ≥ 40
X1 , X 2 ≥ 0
1. Different objective functions can give different solutions
2. There can be multiple objectives
3. They can be conflicting

We need ways of absorbing all these and provide an acceptable solution.


Let X1 and X2 be the quantities produced.

Maximize 800X1 + 900X2


Maximize 70X1 + 100X2
Point 800X1 + 900X2 70X1 + 100X2
4X1 + 5X2 ≤ 200 (O1) (O2)
6X1 + 3X2 ≤ 210
(25, 20) 38000 3750
X1 + X2 ≥ 40
(30, 10) 33000 3100
X1 , X 2 ≥ 0
(0, 40) 36000 4000
(20,20) 34000 3400
(10,32) 36800 3900

Which is the solution?


Weighted objective function

Z = α O1 + β O2

Use α = 1 and β = 5 Use α = 1 and β = 10

Z = 1150X1 + 1400X2 Z = 1500X1 + 1900X2


(25, 20) Z = 56750 (25, 20) Z = 75500
(30, 10) Z = 48500 (30, 10) Z = 64000
(0, 40) Z = 56000 (0, 40) Z = 76000

How do we get the weights?

Z is no indication of the two individual objective function values


Point 800X1 + 900X2 70X1 + 100X2
(O1) (O2)
(25, 20) 38000 3750
(30, 10) 33000 3100
(0, 40) 36000 4000
(27, 14) 34200 3290

X = (25, 20). X= (0, 40) are good solutions (Pareto optimal,


non-dominated, efficient solutions)

A noninferior solution is a feasible solution to the problem, x0 ∈ X,


such that no other feasible solution, x, exists for which fp(x) > fp(x0)
for some p = 1, 2, . . . , q, and fi(x) ≥ fi(x0) for all i ≠ p

Can we get one or more of these solutions?


Maximize 800X1 + 900X2
Maximize 70X1 + 100X2

6X1 + 3X2 ≤ 210

Point 800X1 + 900X2 (O1) 70X1 + 100X2 (O2)


(25, 20) 38000 3750
(0, 40) (0, 40) 36000 4000
Z = 36000 (5,36) 36400 3950
(10,32) 36800 3900
(25, 20)
(15,28) 37200 3850
X1 + X2 ≥ 40

(30, 10) Which of these are non dominated


solutions? How many?

4X1 + 5X2 ≤ 200 All points joining (0, 40) and (25, 20)
are non dominated solutions
Targets and Goals
6X1 + 3X2 ≤ 210
4X1 + 5X2 ≤ 200
6X1 + 3X2 ≤ 210 8X1 + 9X2 ≥390
X1 + X2 ≥ 40 (0, 40)
X1 , X 2 ≥ 0 Z = 36000
(25, 20)
Goal is to have a profit of 39000 X1 + X2 ≥ 40
Goal is to have no overtime (30, 10)

(50, 0)
Goals (as constraints) (35, 0)
800X1 + 900X2 ≥ 39000 4X1 + 5X2 ≤ 200

50X1 + 70X2 ≤ 2400


Infeasible
Do we declare it as Infeasible or try and give a solution?
Goal Programming – (Charnes and Cooper)

Goal Programming(GP) is based upon minimizing the sum of the


weighted absolute deviations of objectives fi(x) from targets Ti.

System constraints
4X1 + 5X2 ≤ 200 Minimize w1e1 + w2d2
6X1 + 3X2 ≤ 210
Minimize 2e1 + d2
X1 + X2 ≥ 40
(25, 20) e1 =1000, d2 = 250 Z
X1 , X 2 ≥ 0
Goal constraints = 2250
800X1 + 900X2 – d1 + e1 = 39000
Minimize e1 + 15d2
50X1 + 70X2 – d2 + e2 = 2400
(27.77, 14.44 e1 =3777.77,
d2 = 0 Z = 3777.78
Ranking the goals – Preemptive (Lexicographic) Goal Programming

Priorities are very clear Solve for Goal 1


Goal 1 – Profit goal Minimize {e1},
Goal 2 – Time goal Subject to system constraint +
Goal 1
Minimize {e1}, {d2}
4X1 + 5X2 ≤ 200 (25, 20) with e1 = 1000
6X1 + 3X2 ≤ 210 Profit = 38000
X1 + X2 ≥ 40 Goal 1 is not met.
800X1 + 900X2 – d1 + e1 = 39000 Evaluate Goal 2
Time used = 2650
50X1 + 70X2 – d2 + e2 = 2400
d2 = 250
X1 , X 2 , e 1 , e 2 , d 1 , d 2 ≥ 0
Z = {1000}, {250}
Ranking the goals – Preemptive (Lexicographic) Goal Programming

Priorities are very clear Solve for Goal 1


Goal 1 – Time goal Minimize {d2}
Goal 2 – Profit goal Subject to system constraint +
Goal 1 (time)
Minimize {d2}, {e1}
4X1 + 5X2 ≤ 200 (20, 20) with d2 = 0
6X1 + 3X2 ≤ 210 Goal 1 is met.
X1 + X2 ≥ 40 Add Goal 2 and put d2 = 0
800X1 + 900X2 – d1 + e1 = 39000 Minimize e1
50X1 + 70X2 – d2 + e2 = 2400
X1 , X 2 , e 1 , e 2 , d 1 , d 2 ≥ 0 (27.78, 14.44) e1 = 3777.78
Profit = 35222.22; Time = 2400
Z = {0}, {3777.78}
Minimax Goal Programming

Minimize maximum deviation


(25.79, 18.42) with r = 1790
37210, 2579
Minimize r
4X1 + 5X2 ≤ 200
W = 5; (25.17, 19.66) r = 1170
6X1 + 3X2 ≤ 210
W = 20; (26.42, 17.14) r =2438
X1 + X2 ≥ 40
800X1 + 900X2 – d1 + e1 = 39000 (26, 18) r = 3200
50X1 + 70X2 – d2 + e2 = 2400
e1 ≤ r; 10d2 ≤ r
X1 , X 2 , e 1 , e 2 , d 1 , d 2 ≥ 0

You might also like