Multicriteria Decision

Making
Learning Objectives
9.1 Goal Programming
9.2 Graphical Interpretation of Goal Programming
9.3 Computer Solution of Goal Programming Problems
9.4 The Analytical Hierarchy Process
9.5 Scoring Models
Overview 1
• Study of problems with several criteria, i.e., multiple criteria,
instead of a single objective when making a decision.
• Three techniques discussed: goal programming, the
analytical hierarchy process and scoring models.
• Goal programming is a variation of linear programming
considering more than one objective (goals) in the objective
function.
• The analytical hierarchy process develops a score for each
decision alternative based on comparisons of each under
different criteria reflecting the decision makers’ preferences.
• Scoring models are based on a relatively simple weighted
scoring technique.
Learning Objective 9.1

• Goal Programming
Problem Data (1 of 2)
Beaver Creek Pottery Company Example:
Maximize Z  $40 x1  50 x2

subject to:

1x1  2 x2  40 hours of labor

 4 x1  3 x2  120 pounds of clay
 x1, x2  0

Where: x1 = number of bowls produced

x2 = number of mugs produced
Problem Data (2 of 2)
Adding objectives (goals) in order of importance, the
company…
1. …does not want to use fewer than 40 hours of labor per
day.
2. …would like to achieve a satisfactory profit level of
$1,600 per day.
3. …prefers not to keep more than 120 pounds of clay on
hand each day.
4. …would like to minimize the amount of overtime.
Goal Constraint Requirements
• All goal constraints are equalities that include deviational
variables d  and d  .

• A positive deviational variable d +  is the amount by which a goal

level is exceeded.
• A negative deviation variable d 

is the amount by which a goal
level is underachieved.
• At least one or both deviational variables in a goal constraint
must equal zero.
• The objective function seeks to minimize the deviation from
the respective goals in the order of the goal priorities.
Model Formulation Goal Constraints
Labor goal:
x1  2 x2  d1  d1   40   hours days 

Profit goal:
40 x1  50 x2  d 2   d 2   1,600  $ day 

Material goal:

4 x1  3 x2  d 3   d 3   120  lbs of clay day 

Model Formulation Objective Function
1. Labor goals constraint
(priority 1 - less than 40 hours labor; priority 4 - minimum
overtime):
– Minimize P1d1 , P4d1
2. Add profit goal constraint
(priority 2 - achieve profit of $1,600):
– Minimize P1d1 , P2d 2  , P4d1
3. Add material goal constraint
(priority 3 - avoid keeping more than 120 pounds of clay on hand):
– Minimize P1d1 , P2d 2 , P3d3  , P4d1
Model Formulation Complete Model
Complete Goal Programming Model:
Minimize P1d1 , P2d 2  , P3d3  , P4d1

subject to:
x1  2 x2  d1  d1   40         labor 
40 x1  50 x2  d 2   d 2   1,600    profit 
4 x1  3 x2  d3   d3   120  clay 
x1, x2 , d1 , d1 , d 2  , d 2  , d3  , d3   0
Alternative Forms of Goal Constraints (1 of 2)
• Changing fourth-priority goal “limit overtime to 10 hours”
instead of minimizing overtime:
– d1  d 4   d 4   10
 
–  minimize P1d1 , P2d 2  , P3d 3  , P4d 4 

• Addition of a fifth-priority goal- “important to achieve the goal

for mugs”:
–   x1  d 5   30 bowls
–   x2  d 6   20 mugs
     
–   minimize P1d1 , P2d 2 , P3d 3 , P4d 4 , 4P5d 5  5P5d 6

Two or more goals at the same priority level can be assigned

weights to indicate their relative importance.
Alternative Forms of Goal Constraints (2 of 2)
Complete Model with Added New Goals:
Minimize  P1d1 , P2d 2  , P3d 3  , P4d 4  ,4P5d 5   5P5d 6 

subject to:
x1  2 x2  d1  d1  40
40 x1  50 x2  d 2   d 2   1,600
4 x1  3 x2  d 3   d 3   120
d1  d 4   d 4   10
x1  d 5   30
x2  d 6   20
x1, x2 , d1 , d1 , d 2  , d 2  , d3  , d 3  , d 4  , d 4  , d 5  , d 6   0
Learning Objective 9.2

• Graphical Interpretation of Goal Programming

Graphical Interpretation (1 of 6)

Minimize  P1d1 , P2d 2  , P3d3  , P4d1

subject to:
x1  2 x2  d1  d1  40
40 x1  50 x2  d 2   d 2   1,600
4 x1  3 x2  d3   d 3   120
x1, x2 , d1 , d1 , d 2  , d 2  , d3  , d3   0

Figure 9.1 Goal constraints

Graphical Interpretation (2 of 6)
Figure 9.2 The first-priority goal: minimize d1
Graphical Interpretation (3 of 6)

Figure 9.3 The second-priority goal: minimize d 2
Graphical Interpretation (4 of 6)
Figure 9.4 The third-priority goal: minimize d3 
Graphical Interpretation (5 of 6)

Figure 9.5 The fourth-priority goal: minimize d1
Graphical Interpretation (6 of 6)
Goal programming solutions do not always achieve all goals
and they are not optimal; they achieve the best or most
satisfactory solution possible.
Minimize  P1d1 , P2d 2  , P3d3  , P4d1
subject to:
x1  2 x2  d1  d1  40
40 x1  50 x2  d 2   d 2   1,600
4 x1  3 x2  d3   d 3   120
x1, x2 , d1 , d1 , d 2  , d 2  , d 3  , d 3   0

Solution: x1  15 bowls
x2  20 mugs
d1  15 hours
Learning Objective 9.3
• Computer Solution of Goal Programming Problems with
QM for Windows and Excel
Computer Solution Using Excel (1 of 3)
Exhibit 9.4
Computer Solution Using Excel (2 of 3)
Exhibit 9.5
Computer Solution Using Excel (3 of 3)
Exhibit 9.6
Solution for Altered Problem Using Excel
(1 of 6)

Minimize  P1d1 , P2d 2  , P3d 3  , P4d 4  ,4P5d5   5P5d 6 

subject to:
x1  2 x2  d1  d1  40
40 x1  50 x2  d 2   d 2   1,600 This model includes
goals for overtime and
4 x1  3 x2  d3   d3   120
maximum storage levels
d1  d 4   d 4   10 for bowls and mugs.
x1  d5   30
x2  d 6   20
x1, x2 , d1 , d1 , d 2  , d 2  , d3  , d3  , d 4  , d 4  , d5  , d 6   0
Solution for Altered Problem Using Excel
(2 of 6)

Exhibit 9.7
Solution for Altered Problem Using
Excel (3 of 6)
Exhibit 9.8
Solution for Altered Problem Using
Excel (4 of 6)
Exhibit 9.9
Solution for Altered Problem Using
Excel (5 of 6)
Exhibit 9.10
Solution for Altered Problem Using
Excel (6 of 6)
Exhibit 9.11
Learning Objective 9.4

• The Analytical Hierarchy Process

Overview 2
• Method for ranking several decision alternatives and
selecting the best one when the decision maker has
multiple objectives, or criteria, on which to base the
decision.
• The decision maker makes a decision based on how
the alternatives compare according to several criteria.
• The decision maker will select the alternative that best
meets the decision criteria.
• A process for developing a numerical score to rank
each decision alternative based on how well the
alternative meets the decision maker’s criteria.
Example Problem Statement
Southcorp Development Company shopping mall site
selection.
• Three potential sites:
– Atlanta
– Birmingham
– Charlotte.
• Criteria for site comparisons:
– Customer market base.
– Income level
– Infrastructure
– Transportation
Hierarchy Structure
• Top of the hierarchy: the objective (select the best site).
• Second level: how the criteria contribute to the objective.
• Third level: how each of the three alternatives contributes
to each of the criteria.
General Mathematical Process
• Mathematically determine preferences for sites with
respect to each criterion.
• Mathematically determine preferences for criteria (rank
order of importance).
• Combine these two sets of preferences to
mathematically derive a composite score for each site.
• Select the site with the highest score.
Pairwise Comparisons (1 of 2)
• In a pairwise comparison, two alternatives are compared
according to a criterion and one is preferred.
• A preference scale assigns numerical values to different
levels of performance.
Pairwise Comparisons (2 of 2)
Table 9.1 Preference scale for pairwise comparisons

Preference Level Numeric Value

Equally preferred 1
Equally to moderately preferred 2
Moderately preferred 3
Moderately to strongly preferred 4
Strongly preferred 5
Strongly to very strongly preferred 6
Very strongly preferred 7
Very strongly to extremely preferred 8
Extremely preferred 9
Pairwise Comparison Matrix
A pairwise comparison matrix summarizes the pairwise
comparisons for a criteria.
Customer Customer Customer
Site
Market (A) Market (B) Market (C)
A 1 3 2

B one1third 1 one1fifth
3 5
C one1half 5 1
2
Developing Preferences Within Criteria
(1 of 3)

In synthesization, decision alternatives are prioritized within

each criterion

Customer Customer Customer Customer Customer Customer

Site Site
Market (A) Market (B) Market (C) Market (A) Market (B) Market ( C)
A 1 3 2
A start fraction, six
6
over eleven, end
3
three ninths five eights
5
B 1
one third
1 1
one fifth fraction
11 9 8
3 5 B start fraction, 2
2
over 11, end 1
one ninth one sixteenth
1
C start fraction one
1
five ninths start fraction one
1
fraction
11 9 16
half over 11 over
6 end fraction
5 sixteenth over 5 end
fraction
2 9 16
11 C 3
start fraction, 3
over 11, end 5
five ninths five sixteenths
5
5 fraction
11 9 16
6
Developing Preferences Within
Criteria (2 of 3)
The row average values represent the preference vector
Table 9.2 The normalized matrix with row averages

Customer Customer Customer

Site Row Average
Market (A) Market (B) Market (C)
A 0.5455 0.3333 0.6250 0.5012
B 0.1818 0.1111 0.0625 0.1185
C 0.2727 0.5556 0.3125 0.3803
Blank Blank Blank Blank 1.0000
Developing Preferences Within
Criteria (3 of 3)
Preference vectors for other criteria are computed similarly,
resulting in the preference matrix
Table 9.3 Criteria preference matrix

Criterion Criterion Criterion Criterion

Site
(Market) (Income Level) (Infrastructure) (Transportation)
A 0.5012 0.2819 0.1790 0.1561
B 0.1185 0.0598 0.6850 0.6196
C 0.3803 0.6583 0.1360 0.2243
Ranking the Criteria (1 of 2)
Pairwise Comparison Matrix:

Criterion Market Income Infrastructure Transportation

Market 1 one1fifth 3 4
5
Income 5 1 9 7
Infrastructure one1third one1ninth 1 2
3 9
Transportation one 1
fourth 1
one seventh one1half 1
4 7 2
Ranking the Criteria (2 of 2)
Table 9.4 Normalized matrix for criteria with row averages
Criterion Market Income Infrastructure Transportation Row Averages
Market 0.1519 0.1375 0.2222 0.2857 0.1993
Income 0.7595 0.6878 0.6667 0.5000 0.6535
Infrastructure 0.0506 0.0764 0.0741 0.1429 0.0860
Transportation 0.0380 0.0983 0.0370 0.0714 0.0612
Blank Blank Blank Blank Blank 1.0000

Preference Vector for Criteria:

Market 0.1993 
Income 0.6535 
 
Infrastructure 0.0860 
 
Transportation 0.0612 
Developing an Overall Ranking
Overall Score:
Site A score  .1993 .5012   .6535 .2819   .0860 .1790   .0612 .1561  .3091
Site B score  .1993 .1185   .6535 .0598   .0860 .6850   .0612 .6196   .1595
Site C score  .1993 .3803   .6535 .6583   .0860 .1360   .0612 .2243   .5314

Overall Ranking:

Site Score
Charlotte 0.5314
Atlanta 0.3091
Birmingham 0.1595
Blank 1.0000
Summary of Mathematical Steps (1 of 2)
1. Develop a pairwise comparison matrix for each decision
alternative for each criteria.
2. Synthesization
a. Sum each column value of the pairwise comparison
matrices.
b. Divide each value in each column by its column sum.
c. Average the values in each row of the normalized
matrices.
d. Combine the vectors of preferences for each
criterion.
Summary of Mathematical Steps (2 of 2)
3. Develop a pairwise comparison matrix for the criteria.
4. Compute the normalized matrix.
5. Develop the preference vector.
6. Compute an overall score for each decision alternative
7. Rank the decision alternatives.
Learning Objective 9.5

• Scoring Model
Overview 3
Each decision alternative is graded in terms of how well it
satisfies the criterion according to following formula:
Si   g ij w j

where:
wj = a weight between 0 and 1.00 assigned to criterion j;
1.00 important, 0 unimportant;
sum of total weights equals one.
gij = a grade between 0 and 100 indicating how well
alternative i satisfies criteria j;
100 indicates high satisfaction, 0 low satisfaction.
Example Problem (1 of 2)
Mall selection with four alternatives and five criteria:
Grades for Grades for Grades for Grades for
Weight Alternative Alternative Alternative Alternative
Decision Criteria
(0 to 1.00) (0 to 100) (0 to 100) (0 to 100) (0 to 100)
(Mall 1) (Mall 2) (Mall 3) (Mall 4)
School proximity 0.30 40 60 90 60
Median income 0.25 75 80 65 90
Vehicle traffic 0.25 60 90 79 85
Mall quality and 0.10 90 100 80 90
size
Proximity of other 0.10 80 30 50 70
shopping
Example Problem (2 of 2)
S1  .30  40   .25 75   .25 60   .10 90   .10 80     62.75
S2  .30  60   .25 80   .25 90   .10 100   .10 30    73.50
S3  .30 90   .25  65   .25 79   .10 80   .10 50     76.00
S4  .30  60   .25 90   .25 85   .10 90   .10 70   77.75

Mall 4 preferred because of highest score, followed by

malls 3, 2, 1.
Excel Solution
Exhibit 9.16
Goal Programming Example Problem:
Problem Statement
Public relations firm survey interviewer staffing requirements
determination.
• One person can conduct 80 telephone interviews or 40 personal
interviews per day.
• $50/day for telephone interviewer; $70/day for personal interviewer.
• Goals (in priority order):
1. At least 3,000 total interviews conducted.
2. Interviewer conducts only one type of interview each day;
maintain daily budget of $2,500.
3. At least 1,000 interviews should be by telephone.
Formulate and solve a goal programming model to determine number
of interviewers to hire in order to satisfy the goals
Goal Programming Example: Problem
Solution (1 of 2)
Step 1: Model Formulation:
Minimize  P1d1 , P2d 2  , P3d3 

subject to:
80 x1  40 x2  d1  d1  3,000 interviews
50 x1  70 x2  d 2   d 2   $2,500 budget
80 x1  d 3   d 3   1,000 telephone interviews

where:
x1 = number of telephone interviews

x2 = number of personal interviews

Analytical Hierarchy Process Example
Problem: Problem Statement (1 of 2 )
Purchasing decision, three model alternatives, three
decision criteria. Pairwise comparison matrices:

Gear Gear Gear

Price Price Price Bike Action Action Action
Bike
(X) (Y) (Z) (X) (Y) (Z)
X 1 3 6 X 1 one third
1 one seventh
1
Y one third
1 1 2 3 7
3 Y 3 1 one fourth
1
Z one sixth
1
one half
1 1 4
6 2 Z 7 4 1
Analytical Hierarchy Process Example
Problem: Problem Statement (2 of 2 )
Weight/ Weight/ Weight/
Bike Durabilit Durability Durability
y (X) (Y) (Z)
X 1 3 1
Y one third
1 1 one half
1
3 2
Z 1 2 1

Prioritized decision criteria:

Criteria Price Gears Weight
Price 1 3 5
one1third
Gears 1 2
3
one1fifth one1half
Weight 1
5 2
Analytical Hierarchy Process Example
Problem: Problem Solution (1 of 4)
Step 1: Develop normalized matrices and preference
vectors for all the pairwise comparison matrices for criteria.
Bike Price (X) Price (Y) Price (Z) Row Averages
X 0.6667 0.6667 0.6667 0.6667
Y 0.2222 0.2222 0.2222 0.2222
Z 0.1111 0.1111 0.1111 0.1111
Blank Blank Blank Blank 1.0000

Bike Gear Action (X) Gear Action (Y) Gear Action (Z) Row Averages
X 0.0909 0.0625 0.1026 0.0853

Y 0.2727 0.1875 0.1795 0.2132

Z 0.6364 0.7500 0.7179 0.7014

Blank Blank Blank Blank 1.0000

Analytical Hierarchy Process Example
Problem: Problem Solution (2 of 4)
Weight/ Weight/ Weight/
Row
Bike Durabilit Durability Durability
Averages
y (X) (Y) (Z)
X 0.4286 0.5000 0.4000 0.4429
Y 0.1429 0.1667 0.2000 0.1698
Z 0.4286 0.3333 0.4000 0.3873
Blank Blank Blank Blank 1.0000

Criteria Criteria Criteria

Bike
(Price) (Gears) (Weight)
X 0.6667 0.0853 0.4429
Preference
Y 0.2222 0.2132 0.1698 vectors are
Z 0.1111 0.7014 0.3873 summarized
Analytical Hierarchy Process Example
Problem: Problem Solution (3 of 4)
Step 2: Rank the criteria.

Row
Criteria Price Gears Weight
Averages
Price 0.6522 0.6667 0.6250 0.6479
Gears 0.2174 0.2222 0.2500 0.2299
Weight 0.1304 0.1111 0.1250 0.1222
Blank Blank Blank Blank 1.0000

Price 0.6479 
Gears 0.2299 
 
Weight 0.1222 
Analytical Hierarchy Process Example
Problem: Problem Solution (4 of 4)
Step 3: Develop an overall ranking.

Bike X 0.6667 0.0853 0.4429  0.6479 

   
Bike Y 0.2222 0.2132 0.1698   0.2299 
Bike Z  0.1111 0.7014 0.3837  0.1222 
   

Bike X score  .6667 .6479   .0853 .2299   .4429 .1222   .5057

Bike Y score  .2222 .6479   .2132 .2299   .1698 .1222   .2138
Bike Z score  .1111.6479   .7014 .2299   .3873 .1222   .2806

Overall ranking of bikes: X first followed by Z and Y (sum of

scores equal 1.0000).