Goal Programming
Debabrata Ghosh
The Dewright Company
The Dewright Company is one of the largest producers of power tools in the
United States.
The company is preparing to replace its current product line with the next
generation of productsthree new power tools.
Management needs to determine the mix of the companys three new products
to best meet the following three goals:
1. Achieve a total profit (net present value) of at least $125 million.
2. Maintain the current employment level of 4,000 employees.
3. Hold the capital investment down to no more than $55 million.
Penalty Weights
Goal
Factor
Penalty Weight for Missing Goal
Total profit
5 (per $1 million under the goal)
Employment level
4 (per 100 employees under the goal)
2 (per 100 employees over the goal)
Capital investment
3 (per $1 million over the goal)
Data for Contribution to the Goals
Unit Contribution of Product
Factor
Goal
Total profit (millions of dollars)
12
15
125
Employment level (hundreds of employees)
= 40
Capital investment (millions of dollars)
55
Weighted Goal Programming
A common characteristic of many management science models (linear
programming, integer programming, nonlinear programming) is that they have
a single objective function.
It is not always possible to fit all managerial objectives into a single objective
function. Managerial objectives might include:
Maintain stable profits.
Increase market share.
Diversify the product line.
Maintain stable prices.
Improve worker morale.
Maintain family control of the business.
Increase company prestige.
Weighted goal programming provides a way of striving toward several
objectives simultaneously.
Weighted Goal Programming
With weighted goal programming, the objective is to
Minimize W = weighted sum of deviations from the goals.
The weights are the penalty weights for missing the goal.
Introduce new changing cells, Amount Over and Amount Under, that will
measure how much the current solution is over or under each goal.
The Amount Over and Amount Under changing cells are forced to maintain
the correct value with the following constraints:
Level Achieved Amount Over + Amount Under = Goal
Weighted Goal Programming Formulation for
the Dewright Co. Problem
Let
Pi = Number of units of product i to produce per day (i = 1, 2, 3),
Under Goal i = Amount under goal i (i = 1, 2, 3),
Over Goal i = Amount over goal i (i = 1, 2, 3),
Minimize W = 5(Under Goal 1) + 2Over Goal 2) + 4 (Under Goal 2) + 3 (Over Goal 3)
subject to
Level Achieved
Deviations
Goal
Goal 1: 12P1 + 9P2 + 15P3
(Over Goal 1) + (Under Goal 1) = 125
Goal 2: 5P1 + 3P2 + 4P3
(Over Goal 2) + (Under Goal 2) =
40
Goal 3: 5P1 + 7P2 + 8P3
(Over Goal 3) + (Under Goal 3) =
55
and
Pi 0, Under Goal i 0, Over Goal i 0 (i = 1, 2, 3)
Weighted Goal Programming Spreadsheet
B
3
4
5
6
7
8
9
10
11
12
13
14
15
Goal 1 (Profit)
Goal 2 (Employment)
Goal 3 (Investment)
Units Produced
Contribution per Unit Produced
Product 1
Product 2
Product 3
12
9
15
5
3
4
5
7
8
Product 1
8.33333333
Product 2
0
Product 3
1.66666667
G
Goals
Level
Achieved
125
>=
48.333333 =
55
<=
Goal
125
40
55
Penalty
Weights
Profit
Employment
Investment
J
K
Deviations
Amount Amount
Over
Under
0
0
8.333333
0
0
0
Over
Goal
2
3
Under
Goal
5
4
M
Constraints
Balance
(Level - Over + Under)
125
40
55
Weighted Sum
of Deviations
16.66666667
=
=
=
Goal
125
40
55
Weighted vs. Preemptive Goal Programming
Weighted goal programming is designed for problems where all the goals are
quite important, with only modest differences in importance that can be
measured by assigning weights to the goals.
Preemptive goal programming is used when there are major differences in
the importance of the goals.
The goals are liested in the order of their importance.
It begins by focusing solely on the most important goal.
It next does the same for the second most important goal (as is possible without
hurting the first goal).
It continues the the following goals (as is possible without hurting the previous
more important goals).
Preemptive Goal Programming
Introduce new changing cells, Amount Over and Amount Under, that will
measure how much the current solution is over or under each goal.
The Amount Over and Amount Under changing cells are forced to maintain
the correct value with the following constraints:
Level Achieved Amount Over + Amount Under = Goal
Start with the objective of achieving the first goal (or coming as close as
possible):
Minimize (Amount Over/Under Goal 1)
Continue with the next goal, but constrain the previous goals to not get any
worse:
Minimize (Amount Over/Under Goal 2)
subject to
Amount Over/Under Goal 1 = (amount achieved in previous step)
Repeat the previous step for all succeeding goals.
Preemptive Goal Programming for Dewright
The goals in the order of importance are:
1.
2.
3.
4.
Achieve a total profit (net present value) of at least $125 million.
Avoid decreasing the employment level below 4,000 employees.
Hold the capital investment down to no more than $55 million.
Avoid increasing the employment level above 4,000 employees.
Start with the objective of achieving the first goal (or coming as close as
possible):
Minimize (Under Goal 1)
Then, if for example goal 1 is achieved (i.e., Under Goal 1 = 0), then
Minimize (Under Goal 2)
subject to
(Under Goal 1) = 0
Preemptive Goal Programming Formulation for
the Dewright Co. Problem (Step 1)
Let
Pi = Number of units of product i to produce per day (i = 1, 2, 3),
Under Goal i = Amount under goal i (i = 1, 2, 3),
Over Goal i = Amount over goal i (i = 1, 2, 3),
Minimize (Under Goal 1)
subject to
Level Achieved
Goal 1: 12P1 + 9P2 + 15P3
Goal 2: 5P1 + 3P2 + 4P3
Goal 3: 5P1 + 7P2 + 8P3
Deviations
(Over Goal 1) + (Under Goal 1) =
(Over Goal 2) + (Under Goal 2) =
(Over Goal 3) + (Under Goal 3) =
and
Pi 0, Under Goal i 0, Over Goal i 0 (i = 1, 2, 3)
Goal
125
40
55
Preemptive Goal Programming Formulation for
the Dewright Co. Problem (Step 2)
Let
Pi = Number of units of product i to produce per day (i = 1, 2, 3),
Under Goal i = Amount under goal i (i = 1, 2, 3),
Over Goal i = Amount over goal i (i = 1, 2, 3),
Minimize (Under Goal 2)
subject to
Level Achieved
Goal 1: 12P1 + 9P2 + 15P3
Goal 2: 5P1 + 3P2 + 4P3
Goal 3: 5P1 + 7P2 + 8P3
Deviations
(Over Goal 1) + (Under Goal 1) =
(Over Goal 2) + (Under Goal 2) =
(Over Goal 3) + (Under Goal 3) =
(Under Goal 1) = (Level Achieved in Step 1)
and
Pi 0, Under Goal i 0, Over Goal i 0 (i = 1, 2, 3)
Goal
125
40
55
Preemptive Goal Programming Formulation for
the Dewright Co. Problem (Step 3)
Let
Pi = Number of units of product i to produce per day (i = 1, 2, 3),
Under Goal i = Amount under goal i (i = 1, 2, 3),
Over Goal i = Amount over goal i (i = 1, 2, 3),
Minimize (Over Goal 3)
subject to
Level Achieved
Goal 1: 12P1 + 9P2 + 15P3
Goal 2: 5P1 + 3P2 + 4P3
Goal 3: 5P1 + 7P2 + 8P3
Deviations
(Over Goal 1) + (Under Goal 1) =
(Over Goal 2) + (Under Goal 2) =
(Over Goal 3) + (Under Goal 3) =
(Under Goal 1) = (Level Achieved in Step 1)
(Under Goal 2) = (Level Achieved in Step 2)
and
Pi 0, Under Goal i 0, Over Goal i 0 (i = 1, 2, 3)
Goal
125
40
55
Preemptive Goal Programming Formulation for
the Dewright Co. Problem (Step 4)
Let
Pi = Number of units of product i to produce per day (i = 1, 2, 3),
Under Goal i = Amount under goal i (i = 1, 2, 3),
Over Goal i = Amount over goal i (i = 1, 2, 3),
Minimize (Over Goal 2)
subject to
Level Achieved
Goal 1: 12P1 + 9P2 + 15P3
Goal 2: 5P1 + 3P2 + 4P3
Goal 3: 5P1 + 7P2 + 8P3
Deviations
(Over Goal 1) + (Under Goal 1) =
(Over Goal 2) + (Under Goal 2) =
(Over Goal 3) + (Under Goal 3) =
(Under Goal 1) = (Level Achieved in Step 1)
(Under Goal 2) = (Level Achieved in Step 2)
(Over Goal 3) = (Level Achieved in Step 3)
and
Pi 0, Under Goal i 0, Over Goal i 0 (i = 1, 2, 3)
Goal
125
40
55
Preemptive Goal Programming Spreadsheet
Step 1: Minimize (Under Goal 1)
A
1
2
3
4
5
6
7
8
9
10
11
12
Constraints
Balance
(Level - Over + Under)
125
40
55
=
=
=
Goal
125
40
55
Dewright Co. Goal Programming (Preemptive Priority 1: Minimize Under Goal 1)
Goals
Goal 1 (Profit)
Goal 2 (Employment)
Goal 3 (Investment)
Contribution per Unit Produced
Product 1
Product 2
Product 3
12
9
15
5
3
4
5
7
8
Units Produced
Product 1
3.7037
Level
Achieved
125
40
61.481
>=
=
<=
Goal
125
40
55
Deviations
Amount
Amount
Over
Under
0
0
0
0
6.481
0
Minimize (Under Goal 1)
Product 2
0
Product 3
5.3704
Preemptive Goal Programming Spreadsheet
Step 3: Minimize (Over Goal 3)
A
1
2
3
4
5
6
7
8
9
10
11
12
Constraints
Balance
(Level - Over + Under)
125
40
55
=
=
=
Goal
125
40
55
Dewright Co. Goal Programming (Preemptive Priority 3: Minimize Over Goal 3)
Goals
Goal 1 (Profit)
Goal 2 (Employment)
Goal 3 (Investment)
Contribution per Unit Produced
Product 1
Product 2
Product 3
12
9
15
5
3
4
5
7
8
Units Produced
Product 1
8.333
Product 2
0
Product 3
1.667
Level
Achieved
125
48.333
55
>=
=
<=
Goal
125
40
55
Deviations
Amount
Amount
Over
Under
0
0
8.333333
0
0
0
Minimize (Over Goal 3)
(Under Goal 1) = 0
(Under Goal 2) = 0
Preemptive Goal Programming Spreadsheet
Step 4: Minimize (Over Goal 2)
A
1
2
3
4
5
6
7
8
9
10
11
12
13
Constraints
Balance
(Level - Over + Under)
125
40
55
=
=
=
Goal
125
40
55
Dewright Co. Goal Programming (Preemptive Priority 4: Minimize Over Goal 2)
Goals
Goal 1 (Profit)
Goal 2 (Employment)
Goal 3 (Investment)
Contribution per Unit Produced
Product 1
Product 2
Product 3
12
9
15
5
3
4
5
7
8
Units Produced
Product 1
8.333
Product 2
0
Product 3
1.667
Level
Achieved
125
48.333
55
>=
=
<=
Goal
125
40
55
Deviations
Amount
Amount
Over
Under
0
0
8.333
0
0
0
Minimize (Over Goal 2)
(Under Goal 1) = 0
(Under Goal 2) = 0
(Over Goal 3) = 0
Multi-Objective Decision Making
Many problems have multiple objectives:
Planning the national budget
save social security, reduce debt, cut taxes, build national defense
Admitting students to college
high SAT or GMAT, high GPA, diversity
Planning an advertising campaign
budget, reach, expenses, target groups
Choosing taxation levels
raise money, minimize tax burden on low-income, minimize flight of business
Planning an investment portfolio
maximize expected earnings, minimize risk
Techniques
Preemptive goal programming
Weighted goal programming
