You are on page 1of 16

OPR PROJECT REPROT

Multi-objective decision making

SUBMITTED BY
ROHAN PALIWAL
B19159
1
Contents
Introduction...............................................................................................................................................2
Constraints.............................................................................................................................................4
Solution of the Objective 1 of Linear Programming Problem................................................................4
Sensitivity Analysis.................................................................................................................................5
Solution of the Objective 2 of Linear Programming Problem................................................................6
Sensitivity Analysis.................................................................................................................................7
Solution of the Objective 3 of Linear Programming Problem................................................................8
Sensitivity Analysis.................................................................................................................................9
Non-compensatory method.................................................................................................................11
Compensatory Method........................................................................................................................13
Interpretation of Non-compensatory and Compensatory Approach...................................................15
From Results for Objective functions....................................................................................................15
Excel Attachment..................................................................................................................................15

2
Introduction
The aim of this project is to understand how the simplex method works in order to arrive at the optimum solution to
the real-life problem faced by a dry fruit merchant with multiple objective functions. In order to achieve this aim,
we have developed a linear programming problem not only to optimize benefit, competitive advantage points but
also to reduce total investment in the form of costs.

Problem
A dry fruit merchant plans to start a business and has to choose between 6 dry fruit types-Almond,
Cashew, Dates, Raisins, Walnut, and Pistachio. On every type of dry fruit, he earns different kinds of
profits. The field where the dry fruit merchant works have a wide number of rivals, with nearly all the
dealers offering the same variants. According to his research, the merchant can gain competitive
advantage, represented in the form of points, which is linked to the quantity and type of dry fruits he
obtains. The following table represents the costs, Profit, and competitive advantage for the various types
of dry fruits:

Types of dry fruit Raisin Pistachio


Almond Cashew Dates s Walnut
Profit 100 200 50 120 400 300
Competitive
Advantage 5 3 4 2 6 8
Cost 800 1000 300 400 1500 1800

Decision Variables
Types of dry Quantity (in kg)
fruit
Almond x1
Cashew x2
Date x3
Raisins x4
Walnut x5
Pistachio x6
Quantity (in Initial Capital Display Space
Types of dry
thousand meter) (in (in
fruit
lakhs) square foot)
Almond x1 4000 0.9
Cashew x2 2000 0.5
Dates x3 9000 0.4
Raisins x4 10000 0.2
Walnut x5 6000 0.6
Pistachio x6 2500 0.7

Quantity (in kg) Floor Space Insurance (in


Types of dry
(in lakhs)
fruit
square foot)
Almond x1 7 0.9
Cashew x2 1 0.5
Dates x3 6 0.1
Raisins x4 6 0.6
Walnut x5 9 0.4
Pistachio x6 4.5 0.3

The objective is to find out the optimum mix of the quantity of each type of dry fruits to be procured
given the above objective functions and following constraints:
Equation 1: Total initial capital for advertisement and dealership = 4000x1 + 2000x2 + 9000x3 +
10000x4 + 6000x5 + 2500x6
Equation 2: Total display space = 0.9x1 + 0.5x2 + 0.4x3 + 0.2x4 + 0.6x5 + 0.7x6
Equation 3: Total floor space = 7x1 + x2 + 6x3 + 6x4 + 9x5 + 4.5x6
Equation 4: Total money spent on insurance = 0.9x1 + 0.5x2 + 0.1x3 + 0.6x4 + 0.4x5 + 0.3x6
Equation 5: x1, x2, x3, x4, x5, x6 >= 0

Objectives
There are the three conflicting objectives that the dry fruits merchant is facing:
1. Maximization of total profit:
Max: 100x1 + 200x2 + 50x3 + 120x4 + 400x5 + 300x6
2. Maximization of competitive advantage:
Max: 5x1 + 3x2 + 4x3 + 2x4 + 6x5 + 8x6
3. Minimization of total cost:
Min: 800x1+ 1000x2 + 300x3 + 400x4 + 1500x5 + 1800x6
Constraints

In order to achieve the above-mentioned contrary objectives, the following constraints cannot be
violated by the dry fruits merchant:

S. No. Description Equation


Maximum initial capital for
4000x1 + 2000x2 + 9000x3 + 10000x4 + 6000x5 +
1 dealership
2500x6 <= 200000
2 Maximum display space 0.9x1 + 0.5x2 + 0.4x3 + 0.2x4 + 0.6x5 + 0.7x6 <=
12
4 Maximum floor space 7x1 + x2 + 6x3 + 6x4 + 9x5 + 4.5x6 <= 120
5 Maximum insurance 0.9x1 + 0.5x2 + 0.1x3 + 0.6x4 + 0.4x5 + 0.3x6 <=
20
6 Minimum display space 0.9x1 + 0.5x2 + 0.4x3 + 0.2x4 + 0.2x5 + 0.7x6 >= 5

Solution of the Objective 1 of Linear Programming Problem

The first objective function is to maximize the total profit:


Max: 100x1 + 200x2 + 50x3 + 120x4 + 400x5 + 300x6

After using the excel solver, the following results were found:

Type of dry fruits Almond Cashew Dates Raisins Walnut Pistachio


Quantity of dry fruits as per
0 9.23077 0 0 12.3077 0.0
OF1

Interpretation:
In order to make the maximum Profit, the dry fruit merchant should procure approximately 9.23 kg of
cashew and 12.3 kg of walnut. In doing so, the total Profit that he will make is INR 6769.23
Sensitivity Analysis:

Objective Allowable Allowable


Cell Name Final Value Reduced Coefficient Increase Decrease
Cost
Quantity of dry fruits
$B$14 as 0 -366.6666667 100 366.6666667 1E+30
per OF1 almond
Quantity of dry fruits
$C$14 as 9.230769231 0 200 133.3333333 47.22222222
per OF1 cashew
Quantity of dry fruits
$D$14 as 0 -216.6666667 50 216.6666667 1E+30
per OF1 dates
Quantity of dry fruits
$E$14 as 0 -74.87179487 120 74.87179487 1E+30
per OF1 raisins
Quantity of dry fruits
$F$14 as 12.30769231 0 400 1400 104.2857143
per OF1 walnuts
Quantity of dry fruits
$G$14 as 0 -43.58974359 300 43.58974359 1E+30
per OF1 pistachio

Shadow Constraint Allowable Allowable


Cell Name Final Value Price R.H. Side Increase Decrease
Constraint 1 >= Solution 2.07692307
7.076923077 0 5 1E+30
7
Constraint 2 <= Solution 92307.69231 0 200000 1E+30 107692.3077
Constraint 3 <= Solution 9.95121951
12 358.974359 12 1.88372093
2
Constraint 4 <= Solution 120 20.51282051 120 40.5 96
Constraint 5 <= Solution 9.538461538 0 20 1E+30 10.46153846
Constraint 6 <= Solution 2.07692307
7.076923077 0 5 1E+30
7

Reduced Cost:
 For dry fruit merchant to procure almond and add dry fruit in its variety without reducing its
profits, he can seek to locate almond from places so that almond profit increases by 366,667. If
that sum increases the Profit on almond, dry fruit merchant will procure almond and receive the
same Profit.
 For dry fruit merchant to procure dates and add in its variety of dry fruit without reducing his
profits, he should try locating dates from places such that Profit on dates increases by 216.67.
Once that amount increases the Profit on dates, dry fruit merchant can procure dates and earn the
same Profit.
 For dry fruit merchant to procure raisins and add dry fruit in its variety without reducing its
profits, he should try to locate raisins from places so that the Profit on raisins increases by 74.87.
When that amount raises the Profit on walnuts, the dry fruit merchant will purchase the raisins
and earn the same Profit.
 For dry fruit merchant to procure pistachio and add dry fruit in its variety without reducing its
profits, he should try to locate pistachio from places such that Profit on pistachio increases by
43.58. Once that amount increases the Profit on pistachio, dry fruit merchant can procure
pistachio and earn the same Profit.

Coefficient of Objective Function:


 Dry fruit merchant will end up obtaining the same quantity of cashew and raisins, even if the
Profit on cashew and raisins increases by 133.33 and 1400. However, the procurement portion of
variants will change over and above the given limit.
 Dry fruit merchant will end up procuring the same quantities of cashew and raisins, even if Profit
on cashew and raisins decreases by 47.2 and 104.28, respectively. However, the procurement
portion of variants will change over and above the given limit.
Resource Constraint:
 If the limit 3 varies (increases or decreases) by unit, the overall Profit varies in the same direction
by 358,97. The linear proportionate increase, however, will only occur until the maximum
increase of 9.95 units and decrease until the maximum decrease of 1.88 units, from which the
change ratio in the optimized benefit feature will vary.
 If the limit 4 varies (increases or decreases) by unit, the total benefit varies by 20.51 in the same
direction. However, the linear proportionate increase will only continue till a maximum increase
of 40.5 units and decrease till a maximum decrease of 96 units, beyond which the proportion of
change in optimized profit function will be different.

Solution of the Objective 2 of Linear Programming Problem

The second objective function is to maximize total competitive advantage:


Max: 5x1 + 3x2 + 4x3 + 2x4 + 6x5 + 8x6

After using the excel solver, the following results were found:

Type of dry fruit Almond Cashew Dates Raisins Walnuts Pistachio


Quantity of dry fruits as per
0 0 0 0 0 17.1
OF2

Interpretation:
In order to earn total competitive advantage, the dry fruit merchant should procure approximately 17.1 kg
of pistachio. In doing so, the total competitive advantage that he will get is 137.143 points.
Sensitivity Analysis:
Objective Allowable Allowable
Cell Name Final Value Reduced Coefficient Increase Decrease
Cost
Quantity of clothes as 0
$B$14 per OF2 Almond -5.285714286 5 5.285714286 1E+30
Quantity of clothes as 0
$C$14 per OF2 Cashew -2.714285714 3 2.714285714 1E+30
Quantity of clothes as
0
$D$14 per OF2 Dates -0.571428571 4 0.571428571 1E+30
Quantity of clothes as 0
$E$14 per OF2 Raisins -0.285714286 2 0.285714286 1E+30
Quantity of clothes as
0
$F$14 per OF2 Walnut -0.857142857 6 0.857142857 1E+30
Quantity of clothes as
$G$14 per OF2 Pistachio 17.14285714 0 8 1E+30 1

Constraint Allowable Allowable


Cell Name Final Value Shadow R.H. Side Increase Decrease
Price
Constraint 1 >= 12 0 5 7 1E+30
Constraint 2 <= 42857.14286 0 200000 1E+30 157142.8571
Constraint 3 <= 12 11.42857143 12 6.666666667 7
Constraint 4 <= 77.14285714 0 120 1E+30 42.85714286
Constraint 5 <= 5.142857143 0 20 1E+30 14.85714286
Constraint 6 >= 12 0 5 7 1E+30

Reduced Cost:
 For dry fruit merchant to procure almond and add dry fruit in its variety without decreasing its
competitive advantage, he will seek to locate almond from places so that the competitive
advantage on almond increases by 5.285. Once that amount increases, the competitive advantage
on almond, dry fruit merchant can procure almond and earn the same competitive advantage.
 For dry fruit merchant to obtain cashew and add dry fruit in its variety without growing its
competitive advantage, he will seek to locate dates from places so that the competitive advantage
on dates increases by 2.714. Once that amount increases, the competitive advantage on cashew,
dry fruit merchant can procure cashew and earn the same competitive advantage.
 For dry fruit merchant to procure dates and add in its variety of dry fruit without reducing his
competitive advantage, he should try locating dates from places such that competitive advantage
on dates increases by 0.571. Once that amount increases the competitive advantage on dates, dry
fruit merchant can procure dates and earn the same competitive advantage.
 For dry fruit merchant to procure raisins and add in its variety of dry fruit without reducing his
competitive advantage, he should try locating raisins from places such that competitive advantage
on raisins increases by 0.285. Once the competitive advantage on raisins increases by that
amount, dry fruit merchant can procure raisins and earn the same competitive advantage.
 For dry fruit merchant to procure raisins and add in its variety of dry fruit without reducing his
competitive advantage, he should try locating raisins from places such that competitive advantage
on raisins increases by 0.285. Once the competitive advantage over raisins increases by this
number, dry fruit merchants will obtain raisins and achieve the same competitive advantage

Coefficient of Objective Function:


 The dry fruit merchant ends up obtaining the same quantities of pistachio, even if the competitive
advantage on pistachio increases by 1E+30. However, the procurement portion of variants will
change over and above the given limit.
 The dry fruit merchant ends up producing the same quantity of pistachio, even though the
competitive advantage on pistachio falls by 1. However, the procurement portion of variants will
change over and above the given limit.
Resource Constraint:
 If the limit 3 varies (increases or decreases) by unit, the total competitive advantage will vary by
11.42 in the same direction. The linear proportionate increase, however, will only continue until
the maximum growth of 6.67 units and decrease until the maximum decrease of 7 units, beyond
which the change ratio in the optimized profit function will differ.

Solution of the Objective 3 of Linear Programming Problem

The third objective function is to minimize total cost:


Min: 800x1+ 1000x2 + 300x3 + 400x4 + 1500x5 + 1800x6

After using the excel solver, the following results were found:

Type of Dry fruit Almond Cashew Dates Raisins Walnuts Pistac


hio
Quantity of dry fruits as per
0 0 12.5 0 0 0.0
OF3

Interpretation:
In order to incur the least total cost, the dry fruit merchant should procure approximately 12.5 kg of dates.
In doing so, the total cost that he will incur is INR 3750.
Sensitivity Analysis:
Objective Allowable Allowable
Cell Name Final Reduced Coefficient Increase Decrease
Value Cost
Quantity of clothes as
$B$15 per OF3 Almond 0 125 800 1E+30 125
Quantity of clothes as
$C$15 per OF3 Cashew 0 625 1000 1E+30 625
Quantity of clothes as
$D$15 per OF3 Dates 12.5 0 300 55.55555 300
Quantity of clothes as
$E$15 per OF3 Raisins 0 250 400 1E+30 250
Quantity of clothes as
$F$15 per OF3 Walnut 0 1350 1500 1E+30 1350
Quantity of clothes as
$G$15 per OF3 Pistachio 0 1275 1800 1E+30 1275

Constraint Allowable Allowable


Cell Name Final Shadow R.H. Side Increase Decrease
Value Price
Constraint 1 >= 5 750 5 3 5
Constraint 2 <= 112500 0 200000 1E+30 87500
Constraint 3 <= 5 0 12 1E+30 7
Constraint 4 <= 75 0 120 1E+30 45
Constraint 5 <= 1.25 0 20 1E+30 18.75
Constraint 6 >= 5 750 5 3 5

Reduced Cost:
 For dry fruit merchant to procure almond and add dry fruit in its variety without increasing its
cost, he should try to locate almond from places such that almond costs decrease by 125. If that
amount reduces the cost of almond, dry fruit merchant can procure almond and incur the same
expense.
 For dry fruit merchant to procure cashew and add dry fruit in its variety without increasing its
cost, he should try to locate cashew from places such that cashew costs decrease by 625. If that
amount reduces the cost of cashew, dry fruit merchant can procure cashew and incur the same
expense.
 For dry fruit merchant to procure raisins and add dry fruit in its variety without raising its cost, he
should try to locate raisins from places so that raisin costs decrease by 250. Once that amount
decreases the cost of raisins, dry fruit merchant can procure raisins and incur the same cost.
 For dry fruit merchant to procure walnuts and add dry fruit in their variety without raising its cost,
he should try to locate walnuts from places such that walnut costs decrease by 1350. Once that
amount decreases the cost of walnuts, dry fruit merchant can procure walnuts and incur the same
cost.
 For dry fruit merchant to procure pistachio and add dry fruit in its variety without increasing its
cost, he should try to locate pistachio from places such that the cost of pistachio decreases by
1275. If pistachio costs decrease by that number, dry fruit merchant may get pistachio and incur
the same cost.

Coefficient of Objective Function:


 The dry fruit merchant ends up purchasing the same sum of dates, even though the dates cost
increases or decreases by 55.55555 and 300, respectively. However, the procurement portion of
variants will change over and above the given limit.
Resource Constraint:

 If the limit 1 varies (increases or decreases) per unit, the overall cost will vary by 750 in the same
direction. The linear proportionate increase, however, will only continue until the maximum
increase of 3 units and decrease until the maximum decrease of 5 units, after which the change
ratio in the optimized cost function will vary.

 If the limit 6 varies (increases or decreases) per unit, the total cost will vary by 750 in the same
direction. The linear proportionate increase, however, will only continue until the maximum
increase of 3 units and decrease until the maximum decrease of 5 units, after which the change
ratio in the optimized cost function will vary.
Solution of the Multi-Objective Linear Programming Problem
We have solved the above problem by two methods:
1. Non-compensatory method
2. Compensatory method
The process followed in each of the two methods, results obtained and its interpretation are explained
below:
Non-compensatory method

Step 1: Used Excel Solver for achieving an optimum solution to each of the given objectives by taking
care of the constraints. Optimum solution was obtained by varying the values of decision variables x1, x2,
x3, x4, x5 and x6 (Used Simplex LP Method)
Following results were obtained:
Decision Variable G2
G1 G3
S.No. – Length (in Competitive
Revenue Max Cost Min
thousand meters) Advantage Max
1 X1 0 0 0
2 X2 9.230769231 0 0
3 X3 0 0 12.5
4 X4 0 0 0
5 X5 12.30769231 0 0
6 X6 0.0 17.1 0.0

G2
G1 G3
S.No. Parameter Competitive
Revenue Max Cost Min
Advantage Max
1 Total Profit 6769.23 5142.86 625
Total Competitive
2
Advantage 101.538 137.143 50
3 Total Cost 27692.3 30857.1 3750

S.No. Parameter Max Min


1 Total Profit - G1 6769.23 625
Total Competitive
2 Advantage – G2 137.143 50
3 Total Cost – G3 30857.1 3750

Step 2: Calculation of µ’s for normalization


µ1 = (G1 – Min) / (Max – Min) =
= ((100x1 + 200x2 + 50x3 + 120x4 + 400x5 + 300x6)– 625) / 6144.23
µ2 = (G2 – Min) / (Max – Min)
= (5x1 + 3x2 + 4x3 + 2x4 + 6x5 + 8x6) – 50) / 87.1429
µ3 = (Max – G3) / (Max – Min)
= (30857.1 - (800x1+ 1000x2 + 300x3 + 400x4 + 1500x5 + 1800x6)) / 27107.1

Step 3: Minimum operator to determine the highest common attainment (HCA) of all
criteria. Additional constraints:
ƛ < = µ1, ƛ < = µ2, ƛ < = µ3, and ƛ<=1
By adding these additional constraints to the existing constraints, the problem is solved with the following
objective: Maximize ƛ
Results:
Following results were obtained by applying the Simplex LP method in EXCEL Solver:

Type of Dry Decision Variables Length (in thousand meters)


fruits
Almond 𝑥1 0
Cashew 𝑥2 0
Date 𝑥3 4.63241
Raisins 𝑥4 0
Walnut 𝑥5 9.33004
Pistachio 𝑥6 1.8

Objective Value
Total Profit 4512.65
Total Competitive Advantage 89.1502
Total Cost 18678.9

Decision Variables Value


µ1 0.63273
µ2 0.44926
µ3 0.44926

Value of ƛ = 0.44926
Compensatory Method

Step 1: Used Excel Solver for achieving an optimum solution to each of the given objectives by taking
care of the constraints. The optimum solution was obtained by varying the values of decision variables x1,
x2, x3, x4, x5 and x6 (Used Simplex LP Method)
Following results were obtained:
Decision Variable G2
G1 G3
S.No. – Quantity (in Competitive
kg) Revenue Max Advantage Max Cost Min
1 X1 0 0 0
2 X2 9.230769231 0 0
3 X3 0 0 12.5
4 X4 0 0 0
5 X5 12.30769231 0 0
6 X6 0.0 17.1 0.0

G2
G1 G3
S.No. Parameter Competitive
Revenue Max Cost Min
Advantage Max
1 Total Profit 6769.23 5142.86 625
Total Competitive
2
Advantage 101.538 137.143 50
3 Total Cost 27692.3 30857.1 3750

S.No. Parameter Max Min


1 Total Profit - G1 6769.23 625
Total Competitive
2
Advantage – G2 137.143 50
3 Total Cost – G3 30857.1 3750

Step 2: Calculation of µ’s for normalization


µ1 = (G1 – Min) / (Max – Min) =
= ((100x1 + 200x2 + 50x3 + 120x4 + 400x5 + 300x6)– 625) / 6144.23
µ2 = (G2 – Min) / (Max – Min)
= (5x1 + 3x2 + 4x3 + 2x4 + 6x5 + 8x6) – 50) / 87.1429
µ3 = (Max – G3) / (Max – Min)
= (30857.1 - (800x1+ 1000x2 + 300x3 + 400x4 + 1500x5 + 1800x6)) / 27107.1
Step 3: Average operator to determine the highest common attainment (HCA) of all criteria using
equal weights to µi’s
Decision Variables Weights
µ1 0.33
µ2 0.33
µ3 0.33

Additional Constraints: µ1, µ2, µ3 >=0


By adding these additional constraints to the existing constraints, the problem is solved with the following
objective: Maximize {(µ1 + µ2 + µ3)/3}
Results:
Following results were obtained by applying Simplex LP method in EXCEL Solver:

Type of Dry Decision Variables Quantity (in kg)


fruits
Almond 𝑥1 0
Cashew 𝑥2 0
Date 𝑥3 0
Raisins 𝑥4 0
Walnut 𝑥5 8.33333
Pistachio 𝑥6 10.0

Objective Value (INR)


Total Profit 6333.33
Total Competitive Advantage 130
Total Cost 30500

Decision Variables Value


µ1 0.92906
µ2 0.91803
µ3 0.01318
Interpretation of Non-compensatory and Compensatory Approach

From Results for Objective functions:


Below is a consolidated report of optimization results using the two techniques:

Compensatory Method
Objective Non compensatory method
(with Equal Weightages)

Total Profit (in thousands) 4512.65 6333.33

Total Competitive Advantage 89.1502 130

Total Cost (in hundreds) 18678.9 30500

It is clear from the table that the compensatory method which averages the objectives gives a much higher
benefit of INR 6333.33 rather than INR 4512.65, which the dry fruit merchant would obtain if he used the
combination suggested by the non-compensatory method. The same result occurs in the case of total
competitive advantage as in the case of non-compensatory method, it is 130 rather than 89,1502. The
higher profit and competitive advantage the compensatory method gives, however, is at the cost of the
higher total cost. In the case of compensatory method, the cost INR 30500 is much higher than in the case
of
In the case of non-compensatory method INR 18678.9.

Excel Attachment

B19159 OPR
Solution.xlsx

You might also like