Goal Programming
Goal programming which reflects
the Simon's theory of satisficing is
widely applied techniques for
modeling modern decision-making
problems.
The advantage of using goal
programming over other techniques is
with dealing with real-world decision
problems is that it reflects the way
manages actually make decisions.
Goal programming allows decision
maker to incorporate environmental,
Goal Programming originally introduced by
A. Charnes and W.W. Cooper and further
developed by Y.Ijiri, S. M. Lee, and others is
similar to linear programming concept.
Goal programming can be employed in
decision problems with a single goal
(objective) and multiple sub goals, as well as
in cases having multiple goals and sub goals.
With in goal programming model, goals may
be achieved only at the expense of other
goals.
The goal programming necessitates
the establishment of a weighting
system for the goals such that lowerranked (or weighted) goals are
considered only after higher-ranked
goals have been satisfied or have
reached the point beyond which no
further improvement is desirable. This
weights can be ordinal or cardinal.
Goal programming is a form of linear
programming, goal programming models
must be formulated under the same
limitations, assumptions, and conditions as
linear programming models (linearity,
divisibility, determinism, etc.).
Goal programming problems can also be
solved by using the simplex method (in a
modified form)
Goal Programming has been widely
applied to decision problems in business
organizations, government agencies, and
nonprofit institutions. Examples include
the following:
Academic administration planning Manpower planning
Accounting analysis
Marketing logistics
Advertising media scheduling
Military strategies
Blood bank logistics
Organizational analysis
Capital budgeting
Personnel administration
Computer resource allocation
Policy analysis
Decision support system planning Portfolio management
Economic policy analysis Production scheduling
Educational system planning
Project management
Energy resources planning
Quality control
Environmental protection Research and development
Facilities layout and location decisions
Transportation logistics
Financial analysis
Urban planning
Health care delivery planning
Water resources planning
Inventory management
Example 13.1 Product Mix
Problem
A manufacturing company
produces three products,
1, 2, and 3. The three
products have resource
requirements as follows:
Labor (hr/unit)
Materials (lb/unit)
Profit ($/unit)
Product 1
Product 2
Product 3
At present the firm has a normal production capacity of 240
hours of labor available daily and a daily supply of 400
pounds of material.
Maximize Z= 3X1+5X2+2X3
SUBJECT TO
5X1+6X2+3X3240
4X1+6X2+3X3400
X1,X2,X30
1.
2.
3.
4.
This model has a single objective, profit
maximization. Now considering the
management developed set of goals, arranged
in order of their importance to the firm.
Because of labor relations difficulties,
management desires to avoid underutilization
of normal production capacity (i.e., no layoffs
of workers).
Management has established a satisfactory
profit level of $500 per day.
Overtime is to be minimized as much as
possible.
Management wants to minimize the purchase
of additional materials because of handling
and storage problems.
The goal constraints developed are as follows:
Labor Utilization
In order to reflect the possibility of underutilization of
labor (as well as overtime), the original linear
programming constraint is reformulated as
5x 2x 4x d d
1
240
The variable
are referred to as
d 1 and d 1
deviational
variables.
variables They represent the
240 d 1 of hours less than
number
(underutilization)
and the number of hours
240d
exceeding
of production
(overtime) for the amount
In the analysis, one of the deviational variable,
or
d1 d1
must always be zero in the solution.
It is not possible to physically have both
underutilization
and over utilization at the
same time.
,d 2
d1
A constraint in which we attempt to minimize
or both is referred to as a goal constraint.
The objective function for underutilization
is specified as follows:
Minimize P 1 d 1
P1 is the preemptive priority designation for
this goal.
pd
1
The term
reflects the d
fact that the first
priority goal of the firm is to minimize
,
the underutilization of labor.
The first goal is to minimize
as close to zero as possible)
( drive it
The minimization of overtime
the fact
d1
that management has ranked third is
reflected in the objective function as follows:
Minimize
Z
pd
1
pd
P3 designates minimization
d 1 of
as the third priority goal.
overtime,
Z represents a multidimensional function
composed of various priority factors and
associated income immensurable objective
criteria.
Profit Level
Managements second goal is to achieve the
satisfactory profit level of $500. This goal constraint
is formulated as
3x 5 x 2x d
1
d 2 500
Where d 2 is underachievement of the profit goal andd 2
is the overachievement of the profit goal. The goal is
reflected in the objective function by minimizing
d2
at
the second priority level.
Minimize
pd
1
pd
2
pd
3
Purchase of Materials
Managements final goal is that daily material
purchases in excess of 400 pounds be minimized.
Formulating, the goal constraint
4 x 1 6 x 2 3 x 3 d 3 d 3 400
where is the over utilization of normal material
d 3 and
requirement
is the purchase of extra materials.
d3
The objective
function at the fourth priority level
MinimizeZ
pd
1
pd
2
pd
3
pd
4
The last term reflects managements desire to minimize
the
Purchase of extra materials at a level of priority below
those of the other three goals.
The goal programming model for the problem can be
summarized as:
MinimizeZ
pd
1
pd
2
pd
3
pd
The last term reflects the managements desire to minimize the
purchase of extra material at a level of priority below those of
the other three goals.
goal programming model can be summarized as follows:
subject to
MinimizeZ
d1
d2
pd
d1
5 x 1 2 x 2 4 x 3 d 1 d 1 240
3 x1 5 x 2 2 x 3 d 2 d 2 500
4 x 1 6 x 2 3 x 3 d 3 d 3 400
x1 , x 2 , x 3 ,d 1 ,d 1 ,d 2 ,d 2 ,d 3 ,d 3 0
Solution of this problem requires that the deviations
from the goals specified in the objective function be
minimized.
d1
The value of the deviational variable associated
with the highest preemptive priority (P1) must be
first minimized to the fullest possible extent.
when no further improvement is possible or desired
for this goal, the value of the deviational variable
(
d2
)
associated with the next highest priority factor, P2 is
minimized, and so on.
The solution procedures is a modified simplex
approach.
Z represents the sum of unattained portions of each
of the goals at different priority levels.
Example 13.2 Weighted
Goals
A small manufacturing firm produces
washers and dryers. Production of either
product requires 1 hour of production
time. The plant has a normal production
capacity of 40 hours per week. A maximum
of 24 washers and 30 dryers can be stored
per week. The profit margin is $80 for a
washer and $40 for a dryer. The manager
has established the following goals,
arranged in order of their priority.
P1: Avoid underutilization of normal
production capacity.
P2: Produce as many washers and dryers
as possible. However, since the profit
margin for a washer is twice that for a
dryer, the manager has twice as much
desire to achieve the production of
washers as to achieve the production
of dryers.
P3: Minimize overtime as much as
possible.
Production Capacity
The first goal constraint reflects the
production time requirements for both
products
x 1 x 2 d 1 . d 1 40
where X1 and X2 are the respective
numbers of washersdand
dryers produced.
1
The deviational variable,
, reflects
d1
underutilization
of the normal production
capacity of 40 hours per week, while
overtime.
MinimizeZ p d p d
Priority goals 1 and 3 can be reflected as
Storage constraint
The production goal constraints are:
x1 d 2 24
x 2 d 3 30
The first goal constraint
represents the
d 3 production goal for
underachievement of the
washers.
The second goal constraint
d 2 is the
underachievement of the production goal for dryers.
The production goals and
have
d 2 d 3 been eliminated,
because these goal levels represent absolute
maximum values (i.e., storage capacities) not to be
exceeded.
This type of constraint is referred to as system
constraint because deviation in the positive
and/or negative direction is prohibited .
Second priority goal is reflected in the
objective function as follows:
MinimizeZ
p d 2p d
p
2
d3
pd
3
The goal programming model is formulated
as:
2
MinimizeZ p 1 d 1
subject to:
x1
x 2 d 1 d 1 40
x 1 d 2 24
x d 30
x , x ,d ,d ,d ,d
2
pd
2
pd
3
Example 13.3 Deviational
Variable Goal Constraint
Extending from the previous problem, the added goal
that overtime not exceed 10 hours per week, if
possible. The priority level of this new goal places
it between the old P1 and P2 levels.
The production goal constraint:
x x d d
1
40
Our new goal is that overtime be restricted to 10
hours, which is formulated as
d d d
10
Another way to formulate the same goal
constraint in terms of decision variables is
adding the allowed overtime of 10 hours to
the original production requirement goal
as follows:
x 1 x 2 d 4 d 4 50
The new
objective
function becomes
MinimizeZ p1 d 1 p 2 d 2 p 3 d
4
pd
3
pd
4
The new second priority goal specifies that
the amount of overtime in excess of 10
hours is to be minimized. This goal is not
incompatible with the goal of minimizing
overtime.
The new goal programming model is
Minimize
Z
pd
1
Subject
x1
pd
x 2 d 1 d 1 40
x 2 d 3 30
d
x ,d ,d
1
to
x 1 d 2 24
d 4 d 4 10
pd
3
pd
4
Example 13.4 Recreational
Facility Funding
A city parks and recreational authority has
been given a federal grant of $600,000 to
expand its public recreational facilities. Four
different types of facilities have been
requested by city council members speaking
for their constitutes: gymnasiums, athletic
fields, tennis courts, and swimming pools.
The total demand by various neighborhoods
has been for 7 gyms, 10 athletic fields, 8
tennis courts, and 12 swimming pools.
Each facility costs a certain amount, requires a
certain number of acres, and has an expected
usage. These parameters are summarized in
the following table:
Facility
Cost ($)
Required
Acres
Expected
Usage
(People/week)
Gymnasium
80,000
1500
Athletic
24,000
3000
Tennis Court
15,000
500
Swimming
Pool
40,000
1000
The park authority has located 50 acres of
land for construction (although more land
could be located if necessary).
The authority has established the
following list of prioritized goals:
P1:The authority must spend the total grant
(otherwise the amount not spent will be
returned to the federal government).
P2: The park authority desires that the
facilities be used weekly by 20,000 or
more people.
P3: If more land is acquired, the additional
amount should be limited to 10 acres.
P4: The authority would like to meet the
demands of the city council members for
new facilities. However, this priority should
be weighted according to the number of
people expected to use each facility.
P5: The park authority wants to avoid
securing land beyond the 50 acres
presently available.
Funding Constraint
The cost requirement for the various facilities are
shown in goal constraint:
80 ,000 x1 24 ,000 x 2 15 ,000 3 40 ,000 x 4 d 1 600 ,000
where X1,X2,X3,X4 are number of facilities of each type to
be constructed.
The deviational variable
is the portion of the grant
d1
not spent.
The deviational variable
d 1 has been eliminated, the
first priority goal is reflected in the objective function
as follows:
Minimize
pd
1
Facility Use
The Expected total weekly usage for all the
facilities is formulated as
1500 x1 3000 x 2 500 x 3 1000 x 4 d 2 d 2 20 ,000
The deviational variables are the amounts of
weekly underutilization or over utilization of
the facilities. The priority 2 goal of
minimizing under utilization is shown in
objective function as
Minimize
pd
1
pd
2
Land Requirements
The land requirements for the various facility types are
reflected in the equation as
4 x 1 8 x 2 3 x 3 5 x 4 d 3 d 3 50
The deviational variables represent the amount by
which the number of acres used is less than 50,
,
d
3
and the excess
above 50 acres, . The park authority
d
3
desires that
the amount of land in excess of 50 acres
be limited,
to 10 acres.
d d d
3
10
This
goal is reflected in the objective function by
d4
minimization
of
at the priority 3 level. This goal and the priority 5 goal
are shown in objective function as
Minimize
pd
1
pd
2
pd
3
pd
5
Facility Demand
The demand for facilities is shown in four
goal constraint.
x
x
x
x
d5 d5 7
1
2
d 6 d 6 10
d7 d7 8
3
d 8 d 8 12
4
Minimize
pd
1
pd
2
p d 3p d 6p d
3
p d 2p d
4
pd
5
The complete goal programming model is
formulated as
Z
Minimize
pd
1
pd
2
p d 3 p d 6 p d
3
p d 2p d
7
pd
80Subject
,000 x 24to:
,000 x 15 ,000 x 40 ,000 x d 600 ,000
1500 x1 3000 x 2 500 x 3 1000 x 4 d 2 d 2 20 ,000
4 x1 8 x 2 3 x 3 5 x 4 d 3 d 3 50
d 3 d 4 d 4 10
x1 d 5 d 5 7
x 2 d 6 d 6 10
x3 d 7 d 7 8
x 4 d 8 d 8 12
x j ,d j ,d j 0
Example 13.5 Multiperiod
Investment Problem
A investment firm has $1,000,000 to invest in
four alternatives: stocks, bonds, savings
certificates, and real estate. The firm wishes to
determine the mix of investments that will
maximize the cash value at the end of 6 years.
Investment opportunities in stocks and bonds are
available at the beginning of each of the next 6
year. Each dollar invested in stocks at the
beginning of each year will return $1.20 ( a profit
of $0.20) 2 years later, which can be immediately
reinvested in any alternative. Each dollar
invested in bonds at the beginning of each year
will return $1.40 3 years later, which can be
reinvested immediately.
Investment opportunities in savings certificates
are available only once, at the beginning of the
second year. Each dollar invested in certificates
at the beginning of the second year will return
$1.80 4 years later. Investment opportunities in
real estate are available at the beginning of the
fifth and sixth years. Each dollar invested in
real estate will return $1.10 a year later.
The management of the firm wishes to
determine the optimal mix of investments in
the various alternatives that will achieve the
following goals, listed in the order of their
importance.
P1:: In order to maximize risk, the total amount
invested in stocks and bonds should be limited
to 40% of the total investment.
P2: The amount invested in savings certificates
should be at least 25% of the total investment.
P3: Real estate is expected to be very attractive in
the future. Thus, management would like to
invest at least $300,000 in real estate.
P4: The total cash value by the end of the sixth year
should be maximized.
Si= amount of money invested in stocks at the
beginning of year i; i=1,2,3,4,5
Bi=amount of money invested in bonds
C2= amount of money invested in saving
certificates in year 2
Ri=amount of money invested in real estate
Ii= amount of money held idle and not invested
during year i;
Decision Variables
Year 1 Year 2 Year 3 Year 4 Year 5 Year 6
S
1
S3
S2
S4
B4
B
1
S
5
B
2
B3
C2
R5
R
6
System constraints
Investment opportunity constraint in first year
Year 1: S1 +B1+I1=1,000,000
I1 is the amount of money not invested at the
beginning of year 1.
Second year, the investment will be S 2,,B2, and
C 2.
Investment Opportunities Amounts Available
Year 2: S2+B2+C2+I2 =I1
Year 3: S3+B3+I3
=I2+1.2S1
Investment Opportunities Amounts
Available
Year 4: S4+B4+I4
=I3+1.2S2+1.4B1
Year 5: S5+R5+I5
=I4+1.2S3+1.4B2
Year 6: R6+I6
=I5+1.2S4+1.4B3+1.8C2+1.1R5
Goal Constraint
We can formulate the four goal constraint as
follows:
S B
P1: The total amount invested in stocks and
bonds,
S B C R
should not exceed 40% of the total
We can formulate following goal constra int
investment in all
the alternatives,
5
i 1
i 1
i 1
i 1
i 1
i 1
i 5
S i Bi d 1 d 1 0.4
Re arranging , Minimize
i 1
i 1
i 5
S 1 Bi C 2 Ri
i 1
i 1
i 5
0.6 S i 0.6 B i 0.4 C 2 0.4 Ri d 1 d 1 0
P2: Since the amount invested in savings
certificates should be at least 25% of the total
investment,
from the
d 2 we should minimize
following goal constraint:
5
i 1
i 1
i 5
0.25 S i 0.25 B i 0.75 C 2 0.25 Ri d 2 d 2 0
P3: For the real estate investment, we should
minimize
in the following goal constraint:
6
Ri d 3 d 3 300 ,000
i 5
P4: Our goal is to maximize the total cash
value by the end of the sixth year. The
investments alternatives are S5,B
4, and R6.
d4
By setting a cash value at an arbitrarily
large number M ($500,000,00) and
minimize
,
we
will
Mmaximizing the
1.2 S 5 1.4 B4 1.1R6 I 6 d 4be
cash value. We can formulate as follows:
The complete goal programming model
can be summarized as:
Zpd
Minimize
subject
pd
2
pd
pd
4
to
S B I 1,000 ,000
S B C I I 0
1.2 S S 1.4 B B I I 0
1.2 S S 1.4 B R I I 0
1.2 S 1.4 B 1.8C 1.1R R I I 0
1
i 1
i 1
i 5
0.6 S i 0.6 B i 0.4 C 2 0.4 Ri d 1 d 1 0
i 1
i 1
i 5
0.25 S i 0.25 B i 0.75 C 2 0.25 Ri d 2 d 2 0
6
R d d
i
i 5
300 ,000
1.2 S 5 1.4 B 4 1.1 R6 I 6 d 4 500 ,000 ,000
S , B ,C , R , I , d , d
i
General Goal Programming
Model
The general goal programming model can be
formulated
as follows:
Minimize Z
p w d w d
k 0 i 1
subject
a x d d
j 1
x ,d ,d
j
ik
ik
to
ij
bi
i 1 ,2 ,.......m
where Pk is the preemptive priority weight
(Pk>>>Pk+1) assigned to goal k (k=0 is reserved for
system
constraint),
and
w
w iknumerical( differential) weights assigned to
ik
are
the
the deviational
variables of goal i at a given priority
d i andk,
d i represent the negative and positive
level
deviations, aij is the technological coefficients of xj
in goal i, and bi, is the ith goal level.