ENGINEERING OPTIMIZATION

Goal Programming

BITS Pilani
Pilani Campus
Outline
• Introduction
• Goal Programming Algorithms
– Weights Method
– Preemptive Method

2
Introduction

3
Goal Programming
• Seeking a compromise solution  minimize
deviation from each goal violation.

• Example: 8.1-1 (Tax planning)

4
Goal Programming
Fairville is a small city with a population of about 20,000 residents.
The annual taxation base for real estate property is $550 million.
The annual taxation bases for food and drugs and for general sales
are $35 million and $55 million, respectively. Annual local gasoline
consumption is estimated at 7.5 million gallons. The city council
wants to develop the tax rates based on four main goals:
1. Tax revenues must be at least $16 million to meet the city’s
financial commitments.
2. Food and drug taxes cannot exceed 10% of all taxes collected.
3. General sales taxes cannot exceed 20% of all taxes collected.
4. Gasoline tax cannot exceed 2 cents per gallon.

Can you write the goals in a mathematical form?

5
Goal Programming
550xp + 35xf + 55xs + .075xg ≥ 16
35xf ≤ .1(550xp + 35xf + 55xs + .075xg )
55xs ≤ .2(550xp + 35xf + 55xs + .075xg )
xg ≤ 2
xp , xf , xs , xg ≥ 0

• xp, xf, xs: tax rates (as proportions of tax bases)

• xg: gasoline tax (cents per gallon)

6
Goal Programming
550xp + 35xf + 55xs + .075xg ≥ 16

55xp − 31.5xf + 5.5xs + .0075xg ≥ 0

110xp + 7xf − 44xs + .015xg ≥ 0

xg ≤ 2

xp , xf , xs , xg ≥ 0

Convert this problem into the standard form.

7
Goal Programming
550xp + 35xf + 55xs + .075xg − S1 + R1 = 16

55xp − 31.5xf + 5.5xs + .0075xg − S2 + R 2 = 0

110xp + 7xf − 44xs + .015xg − S3 + R 3 = 0

x g − S4 + R 4 = 2

xp , xf , xs , xg , S1 , S2 , S3 , S4 , R1 , R 2 , R 3 , R 4 ≥ 0

Can we find a basic feasible solution?

8
Goal Programming
550xp + 35xf + 55xs + .075xg − S1 + R1 = 16

55xp − 31.5xf + 5.5xs + .0075xg − S2 + R 2 = 0

110xp + 7xf − 44xs + .015xg − S3 + R 3 = 0

x g − S4 + R 4 = 2

xp , xf , xs , xg , S1 , S2 , S3 , S4 , R1 , R 2 , R 3 , R 4 ≥ 0

Can we find a basic feasible solution?

Solution: xp = .020, xf = .046, xs=.058, and xg=0
9
Goal Programming
550xp + 35xf + 55xs + .075xg ≥ 16

55xp − 31.5xf + 5.5xs + .0075xg ≥ 0

110xp + 7xf − 44xs + .015xg ≥ 0

xg ≤ 2

xp , xf , xs , xg ≥ 0

• Each inequalities represents a goal of the city council.

• Best may be a compromise solution  each goal may
not be satisfied.
10
Goal Programming
550xp + 35xf + 55xs + .075xg + s1− − s1+ = 16

55xp − 31.5xf + 5.5xs + .0075xg + s2− − s2+ = 0

110xp + 7xf − 44xs + .015xg + s3− − s3+ = 0

xg + s4− − s4+ = 2
Deviational
xp , xf , xs , xg , si− , si+ ≥0 variables

• Deviational variables cannot be basic simultaneously!

• Typically, one zero and one non-zero.

11
Goal Programming
550xp + 35xf + 55xs + .075xg + s1− − s1+ = 16

55xp − 31.5xf + 5.5xs + .0075xg + s2− − s2+ = 0

110xp + 7xf − 44xs + .015xg + s3− − s3+ = 0

xg + s4− − s4+ = 2

xp , xf , xs , xg , si− , si+ ≥ 0

• For the 1st inequality, if s1− ≥ 0 (s1+ = 0 ) then is that

goal satisfied?
12
Goal Programming
550xp + 35xf + 55xs + .075xg + s1− − s1+ = 16

55xp − 31.5xf + 5.5xs + .0075xg + s2− − s2+ = 0

110xp + 7xf − 44xs + .015xg + s3− − s3+ = 0

xg + s4− − s4+ = 2

xp , xf , xs , xg , si− , si+ ≥ 0

• For the 1st inequality, if s1− ≥ 0 (s1+ = 0 ) then is that

goal satisfied? NO
13
Goal Programming
550xp + 35xf + 55xs + .075xg + s1− − s1+ = 16

55xp − 31.5xf + 5.5xs + .0075xg + s2− − s2+ = 0

110xp + 7xf − 44xs + .015xg + s3− − s3+ = 0

xg + s4− − s4+ = 2

xp , xf , xs , xg , si− , si+ ≥ 0

• For the 4th inequality, if s4− ≥ 0 (s4+ = 0 ) then is that

goal satisfied?
14
Goal Programming
550xp + 35xf + 55xs + .075xg + s1− − s1+ = 16

55xp − 31.5xf + 5.5xs + .0075xg + s2− − s2+ = 0

110xp + 7xf − 44xs + .015xg + s3− − s3+ = 0

xg + s4− − s4+ = 2

xp , xf , xs , xg , si− , si+ ≥ 0

• For the 4th inequality, if s4− ≥ 0 (s4+ = 0 ) then is that

goal satisfied? YES
15
Goal Programming
• GP finds a compromise solution  convert each
inequality into a flexible goal  may be violated.

• si− and si+ allow meeting or violating ith goal at will.

• Good compromise solution seeks to minimize the

amount by which each goal is violated.

16
Goal Programming
• Good compromise solution seeks to minimize the
amount by which each goal is violated.

• Can you express a compromise solution for Goal #1

(increase tax revenue) in terms of s1− ?

550xp + 35xf + 55xs + .075xg + s1− − s1+ = 16

17
Goal Programming
• Good compromise solution seeks to minimize the
amount by which each goal is violated.

• Can you express a compromise solution for Goal #1

(increase tax revenue) in terms of s1− ?

550xp + 35xf + 55xs + .075xg + s1− − s1+ = 16

Minimize 𝐬𝟏− (Goal #1)

18
Goal Programming
• Good compromise solution seeks to minimize the
amount by which each goal is violated.

• Can you express a compromise solution for Goal #4

(increase tax revenue) in terms of s1− ?

– Minimize s1− (Goal #1)

– Minimize s2− (Goal #2)
– Minimize s3− (Goal #3)
– Goal #4?
Minimize s4+ (Goal #4)  Why s4+ here?
19
Goal Programming (Example 8.2-1)
TopAd, an advertising agency with 10 employees, has received a contract to
promote a new product. The agency can advertise by radio and television. The
following table gives the number of people reached daily by each type of
advertisement and the cost and labor requirements.

The contract prohibits TopAd from using more than 6 minutes of radio
advertisement. Additionally, radio and television advertisements need to reach at
least 45 million people. TopAd has a budget goal of $100,000 for the project.
How many minutes of radio and television advertisement should TopAd use?
Write constraints and goals in mathematical form.
20
Goal Programming (Example 8.2-1)
Traditional approach
x1 : minutes of radio ads, x2 : minutes of television ads

Maximize: 4x1 + 8x2 (Exposure)

Subject to:
8x1 + 24x2 ≤ 100 (Max. budget)

x1 + 2x2 ≤ 10 (Max. employees)

x1 ≤ 6 (Max. radio time)

Solve using TORA.

21
Goal Programming (Example 8.2-1)
Traditional approach
x1 : minutes of radio ads, x2 : minutes of television ads

Maximize: 4x1 + 8x2 (Exposure)

Subject to:
8x1 + 24x2 ≤ 100 (Max. budget)

x1 + 2x2 ≤ 10 (Max. employees)

x1 ≤ 6 (Max. radio time)

Solution1: 𝑥1 = 5 and 𝑥2 = 2.5 ⇒ Exposure = 40 and Budget = 100

Solution2: 𝑥1 = 6 and 𝑥2 = 2 ⇒ Exposure = 40 and Budget = 96
22
Goal Programming (Example 8.2-1)
Formulate a goal-based approach.

23
Goal Programming (Example 8.2-1)
Goal-based approach
x1 : minutes of radio ads, x2 : minutes of television ads
Constraints:
4x1 + 8x2 ≥ 45 (Exposure goal)

8x1 + 24x2 ≤ 100 (Budget goal)

x1 + 2x2 ≤ 10 (Max. employees)

x1 ≤ 6 (Max. radio time)

Convert to standard form

24
Goal Programming (Example 8.2-1)
Goal-based approach
x1 : minutes of radio ads, x2 : minutes of television ads
Constraints:
4x1 + 8x2 + s1− − s1+ = 45 (Exposure goal)

8x1 + 24x2 + s2− − s2+ = 100 (Budget goal)

x1 + 2x2 ≤ 10 (Max. employees)

x1 ≤ 6 (Max. radio time)

• Minimize 𝑠1− (Goal #1)  minimize underachieving exposure.

• Minimize 𝑠2+ (Goal #2)  minimize overshooting budget.
25
Goal Programming (Example 8.2-1)
Goal-based approach
Minimize:
s1− (Goal #1)
s2+ (Goal #2)
Subject to:
4x1 + 8x2 + s1− − s1+ = 45 (Exposure goal)
8x1 + 24x2 + s2− − s2+ = 100 (Budget goal)
x1 + 2x2 ≤ 10 (Max. employees)
x1 ≤ 6 (Max. radio time)

• Represent multiple goals by a single function:

– Weights method
– Preemptive method
26
Weights Method

27
Goal Programming (Weights method)
• Weights method
– Combined objective constructed as a weighted sum of
individual goals.
• For the TopAds problem
Minimize:
– z = G1 + G2 (Equally important goals)

– z = 2G1 + G2 (G1 more important)

– z = G1 + 2G2 (G2 more important)

28
Goal Programming (Weights method)
• Weights method
– Combined objective constructed as a weighted sum of
individual goals.
• For a general problem with n goals:

• Minimize: 𝐳 = 𝐰𝟏 𝐆𝟏 +𝐰𝟐 𝐆𝟐 + ⋯ + 𝐰𝐧 𝐆𝐧

– wi: positive weights reflecting the relative importance

of each goal.
– specific values of weights  subjective.
29
Goal Programming (Weights method)
Goal-based approach
Minimize:
s1− (Goal #1) and s2+ (Goal #2)
Subject to:
4x1 + 8x2 + s1− − s1+ = 45 (Exposure goal)
8x1 + 24x2 + s2− − s2+ = 100 (Budget goal)
x1 + 2x2 ≤ 10 (Max. employees)
x1 ≤ 6 (Max. radio time)

What would be objective function, if exposure goal (Goal

#1) is twice as important as the budgetary goal (Goal #2)?

30
Goal Programming (Weights method)
Final goal formulation
Minimize: z = 2s1− + s2+
Subject to:
4x1 + 8x2 + s1− − s1+ = 45 (Exposure goal)
8x1 + 24x2 + s2− − s2+ = 100 (Budget goal)
x1 + 2x2 ≤ 10 (Max. employees)
x1 ≤ 6 (Max. radio time)

Solve using TORA.

31
Goal Programming (Weights method)
Final goal formulation
Minimize: z = 2s1− + s2+
Subject to:
4x1 + 8x2 + s1− − s1+ = 45 (Exposure goal)
8x1 + 24x2 + s2− − s2+ = 100 (Budget goal)
x1 + 2x2 ≤ 10 (Max. employees)
x1 ≤ 6 (Max. radio time)

Solution 1: x1 = 5, x2 = 2.5 ⇒ Exposure = 40, Budget = 100

Solution 2: x1 = 6, x2 = 2 ⇒ Exposure = 40, Budget = 96

32
Goal Programming (Weights method) P1
Example: A company is considering three forms of advertising
(see Table) to reach at least 30,000 new customers while spending
no more than $25,000. The company estimates a loss of
$5/customer, if the target of 30,000 is not met. The company also
wants to run 10 television ads, and estimates that not meeting this
target would lead to a brand image loss worth $2000/ad. Formulate
and solve the problem using TORA.

Cost per ad ($) Customers reached

Television 3000 1000
Radio 800 500
Newspaper 250 200

35
Goal Programming (Weights method)
Goals:
G1 = 1000x1 + 500x2 + 200x3 ≥ 30,000
G2 = 3000x1 + 800x2 + 250x3 ≤ 25,000
G3 = x1 ≥ 10
No feasible solution (check with TORA).

Introduce deviational variables:

Constraints
1000x1 + 500x2 + 200x3 + s1− − s1+ = 30,000
3000x1 + 800x2 + 250x3 + s2− − s2+ = 25,000
x1 + s3− − s3+ = 10

Objective function?
36
Goal Programming (Weights method)
Minimize:
z = 5s1− + s2+ + 2000s3−
Introduce deviational variables:
Constraints
1000x1 + 500x2 + 200x3 + s1− − s1+ = 30,000
3000x1 + 800x2 + 250x3 + s2− − s2+ = 25,000
x1 + s3− − s3+ = 10

Solution
x1 = 10, x2 = 0, x3 = 100
z = $30,000 (deviation from goals)

37
Goal Programming (Weights method)
Example: DEWRIGHT COMPANY is considering which mix of new products
should be produced (see Table). The management has established the goals of (1)
achieving a long-run profit of at least $125 million from these products, (2)
maintaining the current employment level of 4,000 employees, and (3) holding
the capital investment to less than $55 million. A discussion led to setting penalty
weights of 5 for missing the profit goal (per $1 million under), 2 for going over
the employment goal (per 100 employees), 4 for going under this same goal, and
3 for exceeding the capital investment goal (per $1 million over). Formulate the
goal programming problem.

38
Goal Programming (Weights method)
Goals:
G1 = 12x1 + 9x2 + 15x3 ≥ 125
G2 = 5x1 + 3x2 + 4x3 = 40
G3 = 5x1 + 7x2 + 8x3 ≤ 55

No feasible solution (check with TORA).

Introduce deviational variables…

39
Goal Programming (Weights method)
Goals:
G1 = 12x1 + 9x2 + 15x3 ≥ 125 (Profit)
G2 = 5x1 + 3x2 + 4x3 = 40 (Employees)
G3 = 5x1 + 7x2 + 8x3 ≤ 55 (Capital)

Introduce deviational variables…

G1 = 12x1 + 9x2 + 15x3 + s1− − s1+ = 125

G2 = 5x1 + 3x2 + 4x3 + s2− − s2+ = 40
G3 = 5x1 + 7x2 + 8x3 + s3− − s3+ = 55

Write the objective function.

40
Goal Programming (Weights method)
Write the objective function.
Minimize:
z =?
Subject to:
G1 = 12x1 + 9x2 + 15x3 + s1− − s1+ = 125
G2 = 5x1 + 3x2 + 4x3 + s2− − s2+ = 40
G3 = 5x1 + 7x2 + 8x3 + s3− − s3+ = 55

41
Goal Programming (Weights method)
Write the objective function.
Minimize:
𝐳 = 𝟓𝐬𝟏− + 𝟐𝐬𝟐+ + 𝟒𝐬𝟐− + 𝟑𝐬𝟑+
Subject to:
G1 = 12x1 + 9x2 + 15x3 + s1− − s1+ = 125
G2 = 5x1 + 3x2 + 4x3 + s2− − s2+ = 40
G3 = 5x1 + 7x2 + 8x3 + s3− − s3+ = 55
xi ≥ 0, si ≥ 0

42
Goal Programming (Weights method)
Write the objective function.
Minimize:
z = 5s1− + 2s2+ + 4s2− + 3s3+
Subject to:
G1 = 12x1 + 9x2 + 15x3 + s1− − s1+ = 125
G2 = 5x1 + 3x2 + 4x3 + s2− − s2+ = 40
G3 = 5x1 + 7x2 + 8x3 + s3− − s3+ = 55
xi ≥ 0, si ≥ 0
Solve using TORA
25 5
x1 = , x2 = 0, x3 =
3 3
+ +
25 −
− −
s1 = s1 = 0; s2 = 0, s2 = ; s3 = 0, s3+ = 0
3
43
Goal Programming (Preemptive method)
SUMMARY
Original goals:
G1 = 12x1 + 9x2 + 15x3 ≥ 125 (profit)
G2 = 5x1 + 3x2 + 4x3 = 40 (employment)
G3 = 5x1 + 7x2 + 8x3 ≤ 55 (investment)

25 5
Solution: x1 = , x2 = 0, x3 =
3 3

1st and 3rd goals fully satisfied, but the employment

level goal of 40 is exceeded by 25/3 (8.33 employees)!

44
Preemptive
Method

45
Goal Programming (Example 8.2-1)
TopAd, an advertising agency with 10 employees, has received a contract to
promote a new product. The agency can advertise by radio and television. The
following table gives the number of people reached daily by each type of
advertisement and the cost and labor requirements.

The contract prohibits TopAd from using more than 6 minutes of radio
advertisement. Additionally, radio and television advertisements need to reach at
least 45 million people. TopAd has a budget goal of $100,000 for the project.
How many minutes of radio and television advertisement should TopAd use?
Write constraints and goals in mathematical form.
46
Goal Programming (Example 8.2-1)
Goal-based approach
x1 : minutes of radio ads, x2 : minutes of television ads
Constraints:
4x1 + 8x2 + s1− − s1+ = 45 (Exposure goal)

8x1 + 24x2 + s2− − s2+ = 100 (Budget goal)

x1 + 2x2 ≤ 10 (Max. employees)

x1 ≤ 6 (Max. radio time)

• Minimize 𝑠1− (Goal #1)  minimize underachieving exposure.

• Minimize 𝑠2+ (Goal #2)  minimize overshooting budget.
47
Goal Programming (Example 8.2-1)
Goal-based approach
Step 0: G1>>>G2
Minimize:
s1− (Goal #1)
s2+ (Goal #2)
Subject to:
4x1 + 8x2 + s1− − s1+ = 45 (Exposure goal)
8x1 + 24x2 + s2− − s2+ = 100 (Budget goal)
x1 + 2x2 ≤ 10 (Max. employees)
x1 ≤ 6 (Max. radio time)

• Represent multiple goals by a single function:

– Preemptive method
48
Goal Programming (Preemptive method)
Final goal formulation
Step 1: Solve LP1
Minimize: z = s1−
Subject to:
4x1 + 8x2 + s1− − s1+ = 45 (Exposure goal)
8x1 + 24x2 + s2− − s2+ = 100 (Budget goal)
x1 + 2x2 ≤ 10 (Max. employees)
x1 ≤ 6 (Max. radio time)

Solve using TORA.

49
Goal Programming (Weights method)
Final goal formulation
Step 2: LP2
Minimize: z = s2+
Subject to:
4x1 + 8x2 − s1+ = 40 (Exposure goal)
8x1 + 24x2 + s2− − s2+ = 100 (Budget goal)
x1 + 2x2 ≤ 10 (Max. employees)
x1 ≤ 6 (Max. radio time)

Solution 1: x1 = 5, x2 = 2.5 ⇒ Exposure = 40, Budget = 100

Solution 2: x1 = 6, x2 = 2 ⇒ Exposure = 40 Budget = 96 50

Goal Programming (Preemptive method)
• Preemptive method
– Ranks goals in order of importance.
– Find optimal solution with respect to the first-priority goals.
– Break ties for optimal solution by considering the second-
priority goals.
– Any remaining ties after this re-optimization broken by
considering the third-priority goals, and so on.

– Sequential and Streamlined procedure.

52
Goal Programming (Preemptive method)
Example: Faced with the unpleasant recommendation to increase company’s
workforce by more than 20%, Dewright Company has reconsidered the original
problem formulation. Consequently, management has concluded that a very high
priority should be placed on avoiding an increase in the workforce. Furthermore,
raising more than $55 million for capital investment for the new products would
be extremely difficult, so a very high priority also should be placed on avoiding
capital investment above this level. Thus, the two goals just discussed should be
the first-priority goals, and the other two original goals should be the second
priority. The relative penalty weights are given in Table.

2M
3M
53
Goal Programming (Preemptive method)
Write the 1st priority goal programming formulation.
Minimize:
z =?
Subject to:
?
?
?

2M
3M
54
Goal Programming (Preemptive method) P2
1st priority goal programming formulation.
Minimize:
z = 2Ms2+ + 3Ms3+
Subject to:
G2 = 5x1 + 3x2 + 4x3 + s2− − s2+ = 40
G3 = 5x1 + 7x2 + 8x3 + s3− − s3+ = 55
xi ≥ 0, si ≥ 0

2M
3M
55
Goal Programming (Preemptive method)
1st priority goal programming formulation.
Minimize:
z = 2Ms2+ + 3Ms3+
Subject to:
G2 = 5x1 + 3x2 + 4x3 + s2− − s2+ = 40
G3 = 5x1 + 7x2 + 8x3 + s3− − s3+ = 55
xi ≥ 0, si ≥ 0

Solve using TORA (or inspection)

s2+ = 0, s3+ = 0, and z = 0
OPTIMIZATION
Innumerable solutions (𝐱 𝟏 , 𝐱 𝟐 , 𝐱 𝟑 ) that satisfy both goals:
FOR
𝐆𝟐 = 𝟓𝐱 𝟏 + 𝟑𝐱 𝟐 + 𝟒𝐱 𝟑 ≤ 𝟒𝟎 2ND-PRIORITY
𝐆𝟑 = 𝟓𝐱 𝟏 + 𝟕𝐱 𝟐 + 𝟖𝐱 𝟑 ≤ 𝟓𝟓 GOALS
56
Goal Programming (Preemptive method)
2nd priority goals shouldn’t deteriorate 1st priority goals.
Minimize:
z =?
Subject to:
?
?
?

2M
3M
57
Goal Programming (Preemptive method) P3
2nd priority goals shouldn’t deteriorate 1st priority goals.
Minimize:
z = 5s1− + 4s2−
Subject to:
G1 = 12x1 + 9x2 + 15x3 + s1− − s1+ = 125
G2 = 5x1 + 3x2 + 4x3 + s2− − 0 = 40
G3 = 5x1 + 7x2 + 8x3 + s3− − 0 = 55
xi ≥ 0, si ≥ 0

2M
3M
58
Goal Programming (Preemptive method)
2nd priority goals shouldn’t deteriorate 1st priority goals.
Minimize:
z = 5s1− + 4s2−
Subject to:
G1 = 12x1 + 9x2 + 15x3 + s1− − s1+ = 125
G2 = 5x1 + 3x2 + 4x3 + s2− = 40
G3 = 5x1 + 7x2 + 8x3 + s3− = 55
xi ≥ 0, si ≥ 0

2M
3M
59
Goal Programming (Preemptive method)
2nd priority goals shouldn’t deteriorate 1st priority goals.
Minimize:
z = 5s1− + 4s2−
Subject to:
G1 = 12x1 + 9x2 + 15x3 + s1− − s1+ = 125
G2 = 5x1 + 3x2 + 4x3 + s2− = 40
G3 = 5x1 + 7x2 + 8x3 + s3− = 55
xi ≥ 0, si ≥ 0
Solve using TORA:
15
x1 = 5, x2 = 0, x3 = UNIQUE
4 SOLUTION
−
35 + (THE END!)
s1 = , s1 = 0; s2− = 0; and s3− = 0
4
60
Goal Programming (Preemptive method)
SUMMARY
Original goals:
G1 = 12x1 + 9x2 + 15x3 ≥ 125 (profit)
G2 = 5x1 + 3x2 + 4x3 = 40 (employment)
G3 = 5x1 + 7x2 + 8x3 ≤ 55 (investment)

15
Solution: x1 = 5, x2 = 0, x3 =
4

Achieves both 1st priority goals and one 2nd priority goals (no
decrease in employment), but falls short of the other 2nd
priority goal (long-run profit 125) by just 35/4!

61
Goal Programming (Preemptive method)
• Streamlined procedure
– finds optimal solution for a preemptive goal programming
problem by solving just one linear programming model.

62
Goal Programming (Preemptive method)
Example: For the Dewright Co. preemptive goal programming problem
summarized in Table. Note:
1. different penalty weights are assigned within each of the two priority levels.
2. individual penalty weights (2 and 3) for the 1st-priority goals are multiplied
by M.

Can you formulate a goal programming problem directly?

2M
3M
63
Goal Programming (Weights method)
Write the objective function.
Minimize:
z = 5s1− + 2Ms2+ + 4s2− + 3Ms3+
Subject to:
G1 = 12x1 + 9x2 + 15x3 + s1− − s1+ = 125
G2 = 5x1 + 3x2 + 4x3 + s2− − s2+ = 40
G3 = 5x1 + 7x2 + 8x3 + s3− − s3+ = 55
xi ≥ 0, si ≥ 0
Solve using TORA
15 SAME
x1 = 5, x2 = 0, x3 = SOLUTION
4 (THE END!)

64
Goal Programming (Preemptive method)
• Streamlined procedure
– finds optimal solution for a preemptive goal programming
problem by solving just one linear programming model.

• More than Two Priority Levels.

– penalty weights for the respective levels now are M1,
M2, . . . , Mp-1, 1.
M1 >> M2 >> M3 >>… Mp-1 >> 1

• Example 8.2-3 (Column-Dropping) [Self study]

65
Example problem (8-3)
The Ozark University admission office is processing freshman applications for
the upcoming academic year. The applications fall into three categories: in-state,
out-of-state, and international. The male–female ratios for in-state and out-of-
state applicants are 1:1 and 3:2, respectively. For international students, the
corresponding ratio is 8:1. The ACT score is an important factor in accepting
new students. The statistics gathered by the university indicate that the average
ACT scores for in-state, out-of-state, and international students are 27, 26, and
23, respectively. The admissions committee has established the following
desirable goals for the new freshman class:
(a) The incoming class is at least 1200 freshmen.
(b) The average ACT score for all incoming students is at least 25.
(c) International students constitute at least 10% of the incoming class.
(d) The female–male ratio is at least 3:4.
(e) Out-of-state students constitute at least 20% of the incoming class.
Formulate the problem as a GP model.
66
Example problem (8-3)
Decision variables:
• x1 : in-state students
• x2 : out-of-state students
• x3 : international students

Goal #1: The incoming class is at least 1200 freshmen.

x1 + x2 + x3 ≥ 1200

67
Example problem (8-3)
Decision variables:
• x1 : in-state, x2 : out-of-state, x3 : international

Goal #1: x1 + x2 + x3 ≥ 1200

Goal #2: The average ACT score is at least 25.
27x1 26x2 23x3
+ + ≥ 25
x1 + x2 + x3 x1 + x2 + x3 x1 + x2 + x3
⇒ 2x1 + x2 − 2x3 ≥ 0

68
Example problem (8-3)
Decision variables:
• x1 : in-state, x2 : out-of-state, x3 : international

Goal #1: x1 + x2 + x3 ≥ 1200

Goal #2: 2x1 + x2 − 2x3 ≥ 0
Goal #3: International students constitute at least 10% of the class.
x3 ≥ 0.1 x1 + x2 + x3
⇒ −0.1x1 − 0.1x2 + 0.9x3 ≥ 0

69
Example problem (8-3)
Decision variables:
• x1 : in-state, x2 : out-of-state, x3 : international

Goal #1: x1 + x2 + x3 ≥ 1200

Goal #2: 2x1 + x2 − 2x3 ≥ 0
Goal #3: −0.1x1 − 0.1x2 + 0.9x3 ≥ 0

Goal #4: The female–male ratio is at least 3:4.

x1 x1
• x1 : in-state  (male) and (female)
2 2 x1 2x2 x3
+ +
• x2 : out-of-state 
3x2
(male) and
2x2
(female) 2 5 9 ≥3
5 5 x1 3x2 8x3 4
8x3 x3 + +
• x3 : international  (male) and (female) 2 5 9
9 9

70
Example problem (8-3)
Decision variables:
• x1 : in-state, x2 : out-of-state, x3 : international

Goal #1: x1 + x2 + x3 ≥ 1200

Goal #2: 2x1 + x2 − 2x3 ≥ 0
Goal #3: −0.1x1 − 0.1x2 + 0.9x3 ≥ 0
x1 x2 5x3
Goal #4: − − ≥0
8 20 9

Goal #5: Out-of-state students at least 20% of incoming class.

x2 ≥ 0.2 x1 + x2 + x3
⇒ −0.2x1 + 0.8x2 − 0.2x3 ≥ 0

71
Example problem (8-3)
Decision variables:
• x1 : in-state, x2 : out-of-state, x3 : international

Goal #1: x1 + x2 + x3 ≥ 1200

Goal #2: 2x1 + x2 − 2x3 ≥ 0
Goal #3: −0.1x1 − 0.1x2 + 0.9x3 ≥ 0
x1 x2 5x3
Goal #4: − − ≥0
8 20 9
Goal #5: −0.2x1 + 0.8x2 − 0.2x3 ≥ 0

All xi ≥ 0

72
Example problem (8-3)
Decision variables:
• x1 : in-state, x2 : out-of-state, x3 : international

Goal #1: x1 + x2 + x3 + s1− − s1+ = 1200

Goal #2: 2x1 + x2 − 2x3 + s2− − s2+ = 0
Goal #3: −0.1x1 − 0.1x2 + 0.9x3 + s3− − s3+ = 0
x1 x2 5x3
Goal #4: − − + s4− − s4+ = 0
8 20 9
Goal #5: −0.2x1 + 0.8x2 − 0.2x3 + s5− − s5+ = 0
All xi , si ≥ 0

Solve using the preemptive method, assuming that the goals are
prioritized in the same order given in the problem.
73
Example problem (8-3)
Goal 1 problem
Minimize 𝑧 = s1−
Subject to :
x1 + x2 + x3 + s1− − s1+ = 1200

Solve using TORA…

Clearly s1− = 0 and multiple solutions are available (by inspection)!

74
Example problem (8-3)
Goal 2 problem
Minimize 𝑧 = s2−
Subject to :
x1 + x2 + x3 + 0 − s1+ = 1200
2x1 + x2 − 2x3 + s2− − s2+ = 0

Solve using TORA…

Clearly s2− = 0 and multiple solutions are available (by inspection)!

75
Example problem (8-3)
Goal 3 problem
Minimize 𝑧 = s3−
Subject to :
x1 + x2 + x3 − s1+ = 1200
2x1 + x2 − 2x3 − s2+ = 0
−0.1x1 − 0.1x2 + 0.9x3 + s3− − s3+ = 0

Solve using TORA…

s3− = 0, x1 = 600, x2 = 600 and multiple solutions are available

(by TORA)!

76
Example problem (8-3)
Goal 4 problem
Minimize 𝑧 = s4−
Subject to :
x1 + x2 + x3 − s1+ ≥ 1200
2x1 + x2 − 2x3 − s2+ ≥ 0
−0.1x1 − 0.1x2 + 0.9x3 − s3+ ≥ 0
x1 x2 5x3
− − + s4− − s4+ ≥ 0
8 20 9
Solve using TORA…

s4− = 0, x1 = 979.6, x2 = 0, x3 = 220.4 and multiple solutions are

available (by TORA)!

77
Example problem (8-3)
Goal 5 problem
Minimize 𝑧 = s5−
Subject to :
x1 + x2 + x3 − s1+ ≥ 1200
2x1 + x2 − 2x3 − s2+ ≥ 0
−0.1x1 − 0.1x2 + 0.9x3 − s3+ ≥ 0
x1 x2 5x3
− − − s4+ ≥ 0
8 20 9
−0.2x1 + 0.8x2 − 0.2x3 + s5− − s5+ ≥ 0

Solve using TORA…

s5− = 0, x1 = 801.3, x2 = 240, x3 = 158.7 and multiple solutions

are still available!
78
Example problem (8-3)
Streamlined approach
Minimize 𝑧 = 108 s1− + 106 s2− + 104 s3− + 102 s4− + 100 s5−
Subject to :
x1 + x2 + x3 + s1− − s1+ = 1200
2x1 + x2 − 2x3 + s2− − s2+ = 0
−0.1x1 − 0.1x2 + 0.9x3 + s3− − s3+ = 0
x1 x2 5x3
− − + s4− − s4+ = 0
8 20 9
−0.2x1 + 0.8x2 − 0.2x3 + s5− − s5+ = 0

Solve using TORA…

79
Thank you!

80