8 MULTI–CRITERIA DECISIONS

8 Multi–Criteria Decisions

A. Goal Programming

108. The management of Suncoast Office Supplies establishes monthly goals, or

quotas, for the types of customers contacted. For the next four weeks, Sun-
coast’s customer contact strategy, calls for the sales-force (which consists of
four sales persons) to do the following: Make 200 contacts with established
customers (who have previously purchased supplied from the firm), and make
120 contacts of new customers. the purpose of this later goal is to ensure that
the sales-force is continuing to investigate new sources of sales.
Making allowances for the travel and waiting time, as well as for the demon-
stration and direct sales time, Suncoast has allocated 2 hours of sales-force
effort to each contact of an established customer. New customer contacts tend
to take longer and require 3 hours per contact. Normally each salesperson
works 40 hours per week, or 160 hours over the 4-week planning horizon. So,
under a normal work schedule, the four salespersons will have 640 (= 4×160)
hours of sales-force time available for customer contacts.
Management is willing to use some overtime, if needed, but is also willing
to accept a solution that uses less than the scheduled 640 hours available.
However management wants both overtime and under-utilization of workforce
limited to no more than 40 hours over the 4-week period. Thus, in terms of
overtime, management’s goal is to use no more than 680 (=640+40) hours of
sales-force time; and in terms of labour utilization, management’s goal is to
use at least 600 (=640-40) hours of salesforce time.
In addition to customer contact goals, Suncoast has established a goal re-
garding sales volume. Based on its experience, Suncoast estimates that each
established customer contacted will generate $250 of sales, and that each new
customer contacted will generate $125 of sales. Management wants to generate
sales revenue of at least $70,000 for the next month.
Given Suncoast’s small salesforce and the short time frame involved, manage-
ment has decided that the overtime goal and the labour utilization goal are
both priority level 1 goals. Management has also concluded that the $70,000
sales revenue goal should be a priority level 2 goal and that the two customer
contact goals should be priority level 3 goals.
Formulate a goal programming problem model for the above.

109. Michigan Motors Corporation (MMC) has introduced a new car. As a part of
its promotional campaign, the marketing department has decided to send per-
sonalized invitation to test drive the new car to two target groups: (1) current
owners of MMC automobile, and, (2) owners of luxury cars manufactured by
one of MMC’s competitors. The cost of sending personalized invitation to each
customer is estimated to be $1 per letter. Based on previous experience with

MSM 2020 35 Joydip Mitra

 8 MULTI–CRITERIA DECISIONS

this type of advertising, MMC estimates that 25% of the customers contacted
from group 1 and 10% from group 2 will test drive the new car. As part of
this campaign, MMC has set the following goals:
(a) Goal 1: Get at least 10,000 customers from group 1 to test drive the new
car.
(b) Goal 2: get at least 5000 customers from group 2 to test drive the new
car.
(c) Goal 3: Limit the expense of sending out invitations to $70,000.
Formulate a goal programming model for the above problem assuming that
goals 1 and 2 are priority level 1 goals and that goal 3 is a priority level 2 goal.
110. In-charge of a committee for promoting a golf tournament, is trying to de-
termine how best to advertise the event during the two weeks prior to the
tournament. The committee obtained the following information about the
three advertising media that they are considering to use. The last column
of the table shows the maximum number of advertisements that can be run
during the next two weeks; these values should be treated as constraints.

Audience reached Cost Max. number of

Category
per advertisement per advertisement advertisement
TV 200,000 $ 2500 10
Radio 50,000 $ 450 15
Newspaper 100,000 $ 500 20

The committee has established the following goals for the campaign:
• Priority Level 1 Goal: Reach at least 4 million people.
• Priority Level 2 Goal: The number of television advertisements should
be at least 30% of the total number of advertisements.
• Priority Level 3 Goal: The number of radio advertisements should not
exceed 20% of the total number of advertisements.
• Priority Level 4 Goal: Limit the total amount spent for advertising to
$20,000.
Formulate a goal programming model for the above problem.
111. The university admissions office is processing freshman applications for the
upcoming academic year. The applications fall in three categories: in-state,
out-of-state, and, international. The male-female ratios for the in-state and
out-of-state applications are 1:1 and 3:2 respectively. For the international
students, the ratio is 8:1. The College Test factor is an important factor
in accepting new students. The statistics gathered by the university indicate
that the average College Test score for in-state, out-of-state, and, international
students are 27, 26, and, 23, respectively. The committee on admissions has
established the following desirable goals for the new freshman class:

MSM 2020 36 Joydip Mitra

 8 MULTI–CRITERIA DECISIONS

• Priority 1: The strength of the incoming class is at least 1200.

• Priority 2: The average College Test score of all incoming students is at
least 25.
• Priority 3: International students constitute at least 10% of the incoming
class.
• Priority 4: The female-male ratio is at least 1:1.
• Priority 5: Out-of-state students constitute at least 20% of the incoming
class.
Formulate the above as a goal programming problem.
112. Mac’s Warehouse is a large discount store that operates 7 days per week. The
store requires the following number of full-time employees working each day
of the week:

Day Sun Mon Tue Wed Thu Fri Sat

No. of Employees 47 22 28 35 34 43 53

Each employee must work for 5 consecutive days each week and then have 2
days off. The store currently has 60 employees. Mac’s priorities for employee
scheduling are as follows:
• Priority 1: The store will like to avoid hiring any additional employee.
• Priority 2: Saturdays and Sundays the store must be fully staffed.
• Priority 3: Fridays must be fully staffed.
• Priority 4: Rest of the days of week also should be fully staffed.
Formulate the above as a goal programming problem to determine the number
of employees who will start their 5-day workweek each day of the week to
achieve the store’s objectives.
113. Mantel produces toy carriage, whose final as- Produced units/run
sembly must include four wheels and two Shift Wheel Seats
seats. The factory producing the parts op- 1 500 300
erates three shifts a day. The adjacent table 2 600 280
provides the number of units produced of each 3 640 360
part in each run in each of the three shifts.
Ideally, the number of wheels produced is exactly twice that of the number of
seats (recall that, each toy will include 4 wheels and 2 seats). However, because
the production rates vary from shift to shift, exact balance in production may
not be possible. Mantel is interested in determining the number of production
runs in each shift that minimizes the imbalance in the production of parts.
The capacity limitations restricts the number of runs to: between 4 and 5 for
shift 1, 10 and 20 for shift 2, and, 3 and 5 for shift 3. Formulate the problem
as a goal programming model.

MSM 2020 37 Joydip Mitra

 8 MULTI–CRITERIA DECISIONS

B. Analytic Hierarchy Process

114. A person wants to buy a car. After test-driving many cars, she has narrowed
her list to three cars A, B, and C. The following data is collected by her.
Decision Alternatives
Characteristics A B C
Price ($) 13,100 11,200 9,500
Colour Black Red Blue
Miles per Gallon (MPG) 19 23 28
Interior Deluxe Above Average Standard
Body Type 4-door midsize 2-door sport 2-door compact
Sound System AM/FM, tape, CD AM/FM AM/FM

She wants to buy a car depending on price, mpg, style, and comfort. How will
she proceed?
(a) Non-numeric data: Not all data are numeric. Style, and Comfort, are
two examples.
(b) Subjectivity: Comparison of decision alternatives is necessarily subjective.
In case of price, a numeric data, the difference between A and C is 3600,
which is (= 100 × (3600/9500)%) more than the price of C. Whether this
difference is significant or not, (or, how significant is the difference) varies
with person. In case of non-numeric data, like style, the subjectivity is
even more obvious.
(c) Finding weights: This being a multi-criteria decision, each criterion is
assigned a non-negative weight (sum of weights is unity). The process is
given below.
i. Pairwise comparisons matrix: Each criterion x, is compared with
each other criterion y, and a subjective judgement is made on which
one is important and by how much, according to the following table.
This decides the entry (x, y) in the pairwise comparison matrix as
given in Table a. below.
Table a. Verbal Response to Numerical Rating
more imp.
Numerical criter.
Verbal judgement rating x y
Equally important 1 1 1
Equally to moderately more important 2 2 1/2
Moderately more important 3 3 1/3
Moderately to Strongly more important 4 4 1/4
Strongly more important 5 5 1/5
Strongly to very strongly more important 6 6 1/6
Very strongly more important 7 7 1/7
Very strongly to extremely more important 8 8 1/8
Extremely more important 9 9 1/9
We present the buyer’s verbal judgement of the pairwise comparison of criteria
in Table b, and corresponding pairwise comparison matrix in Table c below. (In
these tables, we represent Price, Style, and Comfort, by Prc, Sty, and Com.

MSM 2020 38 Joydip Mitra

 8 MULTI–CRITERIA DECISIONS

Table b. Buyer’s subjective comparison

Compare More Imp. How much imp. Value
Prc-Sty Prc Equally to Moderately 2
Prc-Com Prc Equally to Moderately 2
Prc-MPG Prc Moderately 3
Sty-Com Sty Equally to Moderately 2
Sty-MPG Sty Moderately to Strongly 4
Com-MPG Com Moderately to Strongly 4
Table c. Pairwise comparison matrix
Prc Sty Com MPG
Prc 1 2 2 3
Sty 1/2 1 2 4
Com 1/2 1/2 1 4
MPG 1/3 1/4 1/4 1
Sum 7/3 15/4 21/4 12
ii. Normalization: Each entry is divided by sum of the corresponding
column. The weights are computed as row averages. The results are
given in table d below
Table d. Normalized Comparison Table with Weights
Prc Sty Com MPG Weight
Prc 3/7 8/15 8/21 3/12 669/1680 = 0.398
Sty 3/14 4/15 8/21 4/12 502/1680 = 0.299
Com 3/14 2/15 4/21 4/12 366/1680 = 0.218
MPG 1/7 1/15 1/21 1/12 143/1680 = 0.085
Sum 1 1 1 1 1680/1680 = 1.000
iii. Similar pairwise comparisons are given below for cars A, B, and C,
for each of four criteria. Weights of each car with respect to each
criterion can be found as above.
Table e. Pairwise Comparison of A, B, C by each criterion
Prc Sty Com MPG
A 1 1/3 1/4 1 1/3 4 1 2 8 1 1/4 1/6
B 3 1 1/2 3 1 7 1/2 1 6 4 1 1/3
C 4 2 1 1/4 1/7 1 1/8 1/6 1 6 3 1
Sum 8 10/3 7/4 17/4 11/21 8 13/8 19/6 15 11 17/4 3/2

(d) Final computations: The following Table f shows the weights of each of
A, B, and C, with respect to each criterion, and the final priorities of
cars A, B, and C.
Table f. Overall Priority of A, B, and C
Prc Sty Com MPG Overall
Car 0.398 0.299 0.218 0.085 Priority
A 0.123 0.265 0.593 0.087 0.265
B 0.320 0.656 0.341 0.274 0.421
C 0.557 0.080 0.065 0.639 0.314
Here, overall priority of x is computed as
X
P (x) = P (x|y) × P (y)
y

where, y represents criteria Prc, Sty, Com, MPG; x = A, B, C; and,

P (x|y) is priority of x when criterion is y. For example,
P (A) = 0.123 × 0.398 + 0.265 × 0.299 + 0.593 × 0.218 + 0.087 × 0.085 = 0.265.

MSM 2020 39 Joydip Mitra

 8 MULTI–CRITERIA DECISIONS

(e) Consistency Ratio of Pairwise Comparison Matrices:

115. Vacation Location: Students at a college in Delhi are planning for a winter
break vacation to one of three locations: A, B, and C. Decision will be
based on three criteria – weather, cost, and potential fun (based on surveys
of friends). The students have developed the following pairwise comparison
matrices for the locations for each of the above criteria (Table 1), and for the
three criteria (Table 2). Use AHP to decide where they will go for vacation.

Table 1. Table 2.
Pairwise Comparison of A, B, C by each criterion Pairwise Comparison of Criteria
Weather Cost Fun Criterion Weather Cost Fun
A 1 1/3 1/3 1 3 5 1 1/2 5 Weather 1 4 1/4
B 3 1 1 1/3 1 1/2 2 1 3 Cost 1/4 1 1/5
C 3 1 1 1/5 2 1 1/5 1/3 1 Fun 4 5 1

116. The Bay City Parks and Recreation Department is considering building several
new facilities, including a gym, an athletic field, a tennis pavilion, and a pool.
It will base its decision on which facilities to build depending on projected
usage (from surveys) and construction and operating costs. The department
strongly prefers usage to cost as a criterion for selection. The pairwise com-
parison matrices of preferences are given below. Rank the facilities using AHP,
and check the pairwise comparison matrices for consistency.
Pairwise Comparison of Facility by each Criterion
Table 1. Comparison by Usage Table 2. Comparison by Cost
Facility Gym Field Tennis Pool Facility Gym Field Tennis Pool
Gym 1 1/3 3 2 Gym 1 1/4 1/2 3
Field 3 1 5 4 Field 4 1 3 7
Tennis 1/3 1/5 1 1/3 Tennis 2 1/3 1 4
Pool 1/2 1/4 3 1 Pool 1/3 1/7 1/4 1

MSM 2020 40 Joydip Mitra