Sei sulla pagina 1di 44

Goal programming

In this chapter we discuss the Goal programming


technique for solving multi-objective models (linear).
Goal programming problem is a problem of
finding solution which attains a predefined target
for one or more objective function.
If there exists no solution which achieve pre-
specified targets in all objective functions, the
task is to find solution which minimize deviation
from the targets.
If solution with desired target exists, the task of
Goal programming is to identify that particular
solution
Goal Programming Formulation

m goals are expressed as follows:


m
+
ij j i i = bi
a
j =1
x + u o

i = 1, 2,...., m and x j , ui , oi+ 0, for all i, j


+
u , o are deviational variables
i i

representing amount of underachievement and


overachievement of ith goal respectively
Decisions:
The decision maker must analyse each one of the m goals
in terms of whether under or overachievement of the goal
is satisfactory
i) If overachievement of the goal is acceptable , In
objective function we have
Minimize ui-
(underachievement Variable)
ii) If underachievement of the goal is acceptable , In
objective function we have
Minimize oi+
(overachievement Variable)
iii) If exact achievement of the goal is derived ,both ui- and
oi+ must be included in the objective function.
NOTE
In 8th edition following symbols are used
for deviational variable
Underachiement Deviational Variable
Si (Slides Ui - )
Over achievement deviational variable
Si + (Slides Oi + )
Difference between LP and GP
LP identifies from the set of feasible solutions,
the point that optimizes a single objective.
GP determines the point that best satisfies the
set of goals in the decision problem.
GP attempts to minimize the deviations from
the goals.
Example:
Ozark University admissions office is processing
freshman applications for upcoming academic year.
The applications fall into three categories: in-state,
out-of-state and international. The male-female ratio
for in-state, out-of-state and international applicants
are 1:1, 3:2 and 8:1 respectively. The American
College Test (ACT) score is an important factor in
accepting new students. Statistics indicate that average
ACT scores for in-state, out-of-state and international
students are 27, 26 and 23 respectively. The
Committee on admission has established the following
desirable Goals for freshman class:
a. The incoming class has at least 1200 freshman.
b. The average ACT score of all incoming
students is at least 25.
c. International students constitute at least 10% of
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 goal programming
model.
Solution: Let x1, x2 and x3 represent numbers of in-
state, out-of-state and international freshmen.
The goals for the new freshman classes are
expressed as
(a) x1 + x 2 + x3 1200
27 x1 + 26 x 2 + 23x3
(b) 25
x1 + x 2 + x3
x3
(c) 0.1
x1 + x 2 + x3
1/2 x1 + 2 / 5 x 2 + 1 / 9 x3 3
(d)
1 / 2 x1 + 3 / 5 x 2 + 8 / 9 x3 4
x2
(e) 0.2
x1 + x 2 + x3
x1 , x 2 , x3 0

These goal constraints are then simplified as


(a) x1 + x 2 + x3 1200
(b) 2 x1 + x 2 2 x3 0
(c) - 0.1x1 0.1x 2 + 0.9 x3 0
(d) 1 / 8 x1 1 / 20 x 2 5 / 9 x3 0
(e) 0.2 x1 + 0.8 x 2 0.2 x3 0
x1 , x 2 , x3 0
Each of the inequality of the model represent the
goal that the admission committee wants to
establish. So we seek a compromise solution
among these conflicting goal.
The manner in which goal programming finds a
compromise solution is to convert each inequality into
flexible goal in which corresponding constraints may
be violated, if necessary. The flexible goals are
expressed for this model as follows:
+
(a) x1 + x2 + x3 + u o = 1200
1 1
+
(b) 2 x1 + x2 2 x3 + u o = 0
2 2

(c) - 0.1x1 0.1x2 + 0.9 x3 + u3 o3+ = 0


(d) 1 / 8 x1 1 / 20 x2 5 / 9 x3 + u 4 o4+ = 0
(e) 0.2 x1 + 0.8 x2 0.2 x3 + u5 o5+ = 0
x1 , x2 , x3 0, oi+ , ui 0, i = 1,2,3,4,5.
In the given model all constraints are of the type
Thus compromise solution tries to satisfy
following 5 objectives as much as possible
G1: Minimize u1-
G2: Minimize u2-
G3: Minimize u3-
G4: Minimize u4-
G5: Minimize u5-
These functions are minimized subject to
constraints equations of the model
+
x1 + x2 + x3 + u o = 1200
1 1
+
2 x1 + x2 2 x3 + u o = 0
2 2

- 0.1x1 0.1x2 + 0.9 x3 + u3 o3+ = 0


1 / 8 x1 1 / 20 x2 5 / 9 x3 + u4 o4+ = 0
0.2 x1 + 0.8 x2 0.2 x3 + u5 o5+ = 0
x1 , x2 , x3 0, ui , oi+ 0, i = 1,2,3,4,5.
Example 2:
A small paint company manufactures two types of paint, latex and
enamel. In production, the company uses 10 hours of labor to
produce 100 gallons of latex and 15 hours of labor to produce 100
gallons of enamel. The company has 40 hours of daily labor and
30 hours of overtime labor available each week. Each paint
generates a profit at the rate of $1.00 per gallon

(1) avoid the use of overtime


(2) achieve a weekly profit of $1000
(3) produce at least 700 gallons of enamel paint each week

Formulate the problem as goal programming problem.


Rewrite information in a table format

Latex Enamel
Labor for 100 gallons (hours) 10 15
Profit for 1 gallon $1 $1

Remember these are our constraints not our goals.


Goals may not be satisfied (i.e they can go over or under the
required amount) but constraints must always be satisfied.
Decide decision variables
L = no. of 100 gallons of latex paint produced per week

E = no. of 100 gallons of enamel paint produced per week


Write the constraints using your decision variables

Constraint : 10 L + 15 E 70 (Labor constraint)

Write goal constraints using your decision variables

Goal 1: 10 L + 15 E 40 (Avoid Overtime)

Goal 2: 100 L + 100 E 1000 (Profit)

Goal 3: E 7 (Enamel Paint Production)


The flexible goals constraints are expressed for this model as
follows:
10 L + 15 E + u1- - o1+ = 40

100 L + 100 E + u2- - o2+ = 1000

E + u3- - o3+ = 7

Thus compromise solution tries to satisfy following objectives


as much as possible

G1: Minimize o1-


G2: Minimize u2-
G3: Minimize u3-
The Goal programming problem:
G1: Minimize o1-
G2: Minimize u2-
G3: Minimize u3-

Subject to
10 L + 15 E + u1- - o1+ = 40

100 L + 100 E + u2- - o2+ = 1000


E + u3- - o3+ = 7
10 L + 15 E 70
L , E, u1- , o1+ , u2- , o2+ , u3- , o3+ 0
Optimization of a multivariate model with
possibly conflicting goals
Two Methods
 The Weights method
 the Preemptive method
Both methods are based on converting the
multiple objectives into a single function
The Weights Method

Suppose goal programming model has n goals


given as
Minimize Gi; i=1, 2, .., n
The combined objective function used in weighted
method is defined as
Minimize z = w1G1 + w2G2 + .+ wnGn
wi represents positive weights that reflect the
decision makers preference regarding relative
importance of each goal.
wi = 1, for all i signifies that all goals carry equal
weights.
The determination of the specified values of
these weights is subjective.
Example:
In the Ozark University admission situation
described in the first problem, suppose that
the limit on the size of incoming freshman
class must be met, but the remaining
requirement can be treated as a flexible goal.
Further assume that ACT score goal is twice
as important as any of the remaining goals.
(a) Solve the problem & specify whether or
not all goals are satisfied.
Minimize z = 2u2 + u3 + u4 + u5
Subject to x1 + x2 + x3 1200
2x1 + x2 2 x3 + u2 o2+ = 0
-0.1x1 0.1x2 + 0.9 x3 + u3 o3+ = 0
1/ 8 x1 1/ 20 x2 5 / 9 x3 + u4 o4+ = 0
0.2 x1 + 0.8 x2 0.2 x3 + u5 o5+ = 0
x1 , x2 , x3 0, oi+ , ui 0, i = 2,3, 4,5.

Solving by simplex method we get


z = 0 all goals are satisfied


x1 = 801, x2 = 240, x3 = 159 (Approx values)
o2+= 1525.6 ACT score overachieved by
1525.6/1200=1.27 points per student on the

average
o3+= 38.59 number of international
students overachieved by =39 (Approx)
(b) If in addition, the size of the incoming class
can be treated as a flexible goal that is twice as
important as the ACT goal, how would this
change affect the solution?


Minimize z = 4u + 2u + u + u + u
1 2 3 4 5

where the first constraint is expressed as


+
x1 + x2 + x3 + u o = 1200
1 1
Minimize z = 4u1 + 2u2 + u3 + u4 + u5
Subject to x1 + x2 + x3 + u1 o1+ = 1200
+
2x1 + x2 2 x3 + u o = 0
2 2
+
-0.1x1 0.1x2 + 0.9 x3 + u o = 0
3 3
+
1/ 8 x1 1/ 20 x2 5 / 9 x3 + u o = 0
4 4
+
0.2 x1 + 0.8 x2 0.2 x3 + u o = 0
5 5
+
x1 , x2 , x3 0, o , u 0, i = 1, 2,3, 4,5.
i i

Solve by simplex method.


Preemptive Method
Rank goals of the problem in order of
importance
Given an n-goal situation, objectives are
mentioned as
Minimize G1= p1(Highest priority)

.
.
Minimize Gn= pn (Lowest priority)
The variable pi is either Oi+ or Ui- representing
goal i
The solution procedure considers one goal at a
time, starting with the highest priority G1, and
terminating with the lowest Gn.

Step 0: Identify the goals of the model and rank


them in order of priority
G1 = p1 > G2 = p2 > . > Gn = pn
Set i = 1.
Step 1: Solve LPi that minimizes Gi, and let pi = pi*
define the corresponding optimum value of the
deviational variable pi.
If i = n stop; LPn solves the n goal program.
Otherwise, add the constraint pi = pi* to the
constraints of the Gi-problem to ensure that the
value of pi will not be degraded in future
problems.
Set i = i + 1, and repeat step i.
Example: TopAd a new advertising agency with 10
employees has received a contract to promote a new
product. The agency can advertise by radio or television.
The following table provides data about the number of
people reached by each type of advertisement, and the cost
and labor requirements.
Data/min advertisement
Radio Television
Exposure (in millions of persons) 4 8
Cost (in thousands of dollars) 8 24
Assigned employees 1 2
The contract prohibits TopAd from using more than
6 minutes of radio advertisement. Additionally ,
radio and television advertisement need to reach at
least 45 million people. TopAd has set a budget
goal of $1,00,000 for the project. How many
minutes of radio and television advertisement
should TopAd use. Solve the problem by
Preemptive method. Assume that exposure goal
has higher priority.
Solution: Let x1 and x2 be the minutes allocated to
radio & television advertisements. The goal
programming formulation for the problem is
given as follows
Minimize G1 = u1- (Satisfy exposure goal)
Minimize G2 = o2+ (Satisfy budget goal)
subject to
4x1 + 8x2 + u1- - o1+ = 45 (Exposure goal)
8x1 + 24x2 + u2- - o2+ =100 (Budget goal)
x1 + 2x2 10 (Personnel limit)
x1 6 (radio limit)
x1, x2, o1+, u1-, o2+, u2- 0
Step 0: G1 > G2
G1: Minimize u1- (Exposure goal)
G2: Minimize o2+ (Budget goal)
Step 1: Solve LP1.
Minimize G1 = u1-
subject to
4x1 + 8x2 + u1- - o1+ = 45 (Exposure goal)
8x1 + 24x2 +u2- - o2+ = 100 (Budget goal)
x1 + 2x2 10 (Personnel limit)
x1 6 (radio limit)
x1, x2,o1+, u1-, o2+, u2- 0
The optimum solution is x1 = 5 minutes, x2 = 2.5
minutes, u1- = 5 million people, with the
remaining variables equal to zero. The solution
shows that the exposure goal G1 is violated by 5
million persons. In LP1 we have p1 = u1- Thus
the additional constraint we use with the G2-
problem is u1- = 5.
Step 2: We need to solve LP2 whose objective
function is
Minimize G2 = o2+
subject to same set of constraints as in step1
plus additional constraint u1- = 5.
We can solve the new problem.
The additional constraint u1- = 5 can also be
accounted for by substituting out u1- in the first
constraint. The result is that right hand side of the
exposure constraint will be changed from 45 to 40,
thus reducing LP2 to
Minimize G2 = o2+
subject to
4x1 + 8x2 - o1+ = 40 (Exposure goal)
8x1 + 24x2 +u2- - o2+ = 100 (Budget goal)
x1 + 2x2 10 (Personnel limit)
x1 6 (radio limit)
x1, x2, o1+, o2+,u2- 0.
The new formulation is one variable less than
the one in LP1.

Solve.
THE END
Tutorial
(1) A cycle manufacturing company produces two types of
bicycles B1 and B2. The production time for each B1 bicycle is 2
hours and that of B2 is 3 hours. The company have 2000 seats and
2400 tires to be used in both the models in addition 1000 gear
assembly are available to be used only in B2 model. Profit on B1 is
Rs. 300 per unit and on B2 is Rs. 400per unit. The company has
following goals for next month:

(a) Fulfill a contract agreement of 400 B1 bicycles to be delivered


next month.
(b) Achieve at least Rs 100000 profit for the month.
(c) Use no more than 1600 labor hours during the month.
(d) at least 200 tires left over at the end of the month.
Formulate the following problem as a linear goal programming
problem, with a condition the profit goal is twice as important as
the other goals.
(2) You work for an Advertising agency. A customer has identified
three primary target audiences they are trying to reach, and has an
Advertising budget of $ 600,000. They have expressed their target
in the form of three goals:

(a) Ads should be seen by at least 40 million high-income men


(HIM).
(b) Ads should be seen by at least 60 million low-income people
(LIP)
(c) Ads should be seen by at least 35 million high-income women
(HIW)
You recognize that advertising during football games and soap
operas will cover the target audience. The table 1 indicate the
number of viewers from the different categories that will be
viewing these types of programming
Table1:

HIM LIP HIW Cost


Football ad (per min) 7 million 10 million 5 million $ 100,000
Soap Opera ad (per min) 3 million 5 million 4 million $60,000

Suppose each shortfall of one million viewers from the goal


translates to a cost of $ 200,000 for HIM, $ 100,000 for LIP and
$50,000 for HIW. Formulate as a goal programming problem and
hence solve using simplex method.
(3) A company has two machines for manufacturing a product.
Machine 1 make two units per hour, while machine 2 makes three
units per hour. The company has an order of 80 units. Energy
restrictions dictate that only one machine can operate at one time.
The company has 40 hours of regular machine time, but overtime is
available. It costs $ 4 to run machine 1 for one hour, while machine
2 costs $5 per hour.
The company has the following goals
(a) Meet the demand of 80 units exactly
(b) Limit machine overtime to 10 hours
(c) Use the 40 hours of normal machining time
(d) Minimize costs.
Formulate as a goal programming problem.

Potrebbero piacerti anche