HARAMAAYA UNIVERSITY

COLLEGE OF BUSINESS AND ECONOMICS

DEPARTMENT OF ECONOMICS

ENERGY ECONOMICS PROGRAM

DECISION MODELING ASSIGMENT .

#1.(a) A model in decision making

 A model refers to some form of symbolic representation of our assumption about reality.
 A scientific approach to managerial decision making.
 The development of a (mathematical) model of a real-world scenario.
 The process of changing constants into variables ,adding variables, relaxing linear and
other assumptions, including randomness.Typically
 Deterministic Models
Where all the input data value are known with complete certainty
 Probabilistic Models
Where some input data values are uncertain

(b) The importance of a model

 Enhance our understanding of the world to improve our decision making.

 Elaborate a scientific method to solve a problem of duplicable (repeatable) and reduce
bias.
 The model provides insight into the solution of the managerial problem.
 Decisions often must be made in environments that are much more fraughtwith
uncertainty.

Here are a few examples

 A manufacturer introducing a new product into the marketplace.

 A financial firm investing in securities.

 A government contractor bidding on a new contract.

 An agricultural firm selecting the mix of crops and livestock for the upcoming
season.

 An oil company deciding whether to drill for oil in a particular location.

 Decision analysis provides a framework and methodology for rational decision making
when the outcomes are uncertain.
(c) Recursive and non recursive model types

Recursive model

 Recursive model is one where all causal effects are unidirectional and disturbance are
uncorrelated.
 Recursive model is a special case of an equation system where the endogenous variables
are determined one at a time in sequence.
 Recursive models always indentified, simple to estimate.
 A recursive function in general has an extremely high time complexity.
 A recursive function generally has smaller code size.

Non recursive model

 Non recursive contains one or more feedback loops or reciprocal effects.

 Just because a model is indentified does not mean the parameter estimates are correct.
 Consistent estimation of reciprocal paths requires some strict (and often implausible)
assumptions to be met.
 Example, the exogenous variable used to indentify of synchronous parameters must be
meet an instrument variables.
 Non recursive models more flexible
 But can pose problems for indentification.
 Require additional variables for indentification.
 A non recursive model has larger code size and does not has an extremely high time
complexity.

#2.Assumption of a linear programming model

 When we use LP as an approximate representation of a real-life situation, the following

assumptions are inherent:

1. Proportionality:-The contribution of each decision variable to the objective or constraint

is directly proportional to the value of the decision variable.

2. Additivity:-The contribution to the objective function or constraint for any variable is

independent of the values of the other decision variables, and the terms can be added
together sensibly.
 3. Divisibility:-The decision variables are continuous and thus can take on fractional values.

4. Deterministic:-All the parameters (objective function coefficients, right-hand side

coefficients, left-hand side, or technology, coefficients) are known with certainty.

5. Certainty:- We assume that conditions of certainty exists.

• The coefficients in the objective function and constraints are completely known
(deterministic) and do not change during the period being studied.
• For example, profit per unit of product and amount of resources available are
fixed during the planning period.

6. Finiteness:- An optimum solutions cannot be computed in the situations where there are
infinite number of alternative activities and resource restrictions.

7. Optimality:-In linear programming, the maximum profit solutions or the minimum cost
solutions always occur at the corner point of the set of feasible solutions.

#3.About the difference between linear programming and integer programming, whichone
is more efficient and practical in real world?

 Linear programming is basically an iterative technique for optimization in which you

are provided certain data and you have to maximize or minimize the output with
necessary constraints. Basically its a quantitative analysis.

 Integer programming on the other hand is a technique which is used for a qualitative
assessment of the constraints. You might be familiar with sample surveys where you are
asked to put a number for your product satisfaction.

  linear programming can solve problems about minimizing (or maximizing) an

objective function by continuous variables.

 Integer programming allows only integer quantities. We cannot accept fractions.

 Linear programming is used to find optimal solutions to problems using the basis of
a linear function, like a line.
 Integer programming is programming with all variables restricted to integers.

 An integer programming (IP)problem is alinear programming (LP) problem

in which the decision variables are further constrained to take integer values. Both the
objective function and the constraints must be linear. The most commonly used method
for solving an IP is the method of branch-and–bound

 LP problems have a convex set as their feasible region.

 WhileIP problems have set of integer vectors as their feasible region (non-convex).
 A linear program consists of a set of variables, a linear objective function indicating the
contribution of each variable to the desired outcome, and a set of linear constraints
describing the limits on the values of the variables.

Linear programming Problem

 LP improves the quality of decisions .

 The decision-making approach of the user of the technique becomes more objective and
less subjective.
 It also helps in providing better tools for adjustments to meet changing conditions.
 Most business problems involve constraints like raw material availability, etc which must
be taken into consideration. Just coz we can produce so many units does not mean that
they can be sold.
 LP can handle such situation also since it allows modification of its mathematical
solutions.
 Highlighting process the bottle neck in the production process is the most significant
advertizing of this linear programming problem.
 LP treats all relationships among decision variables as linear. However ,generally, neither
the OF nor the constraints in real life situations concerning business and industrial
problems are linearly related to the variables .It may yield fractional valued answer for
decision variables, round the same values may not yield an optimodel does not take into
consideration the effect of time and uncertainty optimal solution.
 Parameters appearing in the model are assumed to be constant but in real life situations,
they are frequently neither known nor constant.
 It deals only with single objectives, whereas in real-life situations we may come across
conflicting multi-objective problems.
 For large problems having many limitations and constraints, the computational
difficulties are enormous, even when the assistance of computer is available. For it, the
main problem can be fragmented into several small problems and solving each one
separately.
 LP helps in attaining the optimum productive resources.

An integer programming

 Real world problems require integer variables

 IP is more efficient
 IP has more modeling power
 More readily implementable answers
 More realistic
 More flexibility
 Easy to program
 But IP
 more complex model
 Much longer to get answer
 More difficult to model
 More difficult to solve

#4.If a basic variable of LPP takes a zero value, discuss the situations occurred in the
model.

 A basic feasible solution is degenerate if at least one of the basic variables is equal to
zero.
 When at least one of the basic variables is zero, then the basic feasible solution to a
Linear Programming Problem is said to be degenerate.
 Degeneracy in a linear programming problem is said to occur when a basic feasible
solution contains a smaller number of non-zero variables than the number of
independent constraints when values of some basic variables are zero and the
Replacement ratio is same.

#5. Discuss the difference between feasible solution, basic solution, and optimal solution in
LPP model.

 A solution (set of values for the decision variables) for which all of the constraints in the
Solver model are satisfied is called a feasible solution.
 A globally optimal solution is one where there are no other feasible solutions with
better objective function values.
 An optimal solution to a linear program is the solution which satisfies all constraints
with maximum or minimum objective function value.
 In simpler words, in a linear programming question we are given an objective
function, some constraints and we have to find minimum or maximum values.
 A basic solution of linear programming problem of a system of m equations and n
variables (m < n) is a solution where at least n-m variables are zero and solving the
remaining equations with unknowns to satisfy certain specified technical conditions.
 A basic solution that happens to have all nonnegative values will be called a basic
feasible solution.

 A nonnegative vector of variables that satisfies the constraints of (P) is called a feasible solution
to the linear programming problem.
 A feasible solution that minimizes the objective function is called an optimal solution.

#6. Discuss about the main methods of solving a linear programming problem.

 I discuss the most important techniques,these are Simplex method and Graphical method

Simplex method
The simplex method is one of the most popular methods to solve linear programming problems.
It is an iterative process to get the feasible optimal solution. In this method, the value of the basic
variable keeps transforming to obtain the maximum value for the objective function. The
algorithm for linear programming simplex method is provided below:
Step 1: Establish a given problem. (i.e.,) write the inequality constraints and objective function.
Step 2: Convert the given inequalities to equations by adding the slack variable to each
inequality expression.
Step 3: Create the initial simplex tableau. Write the objective function at the bottom row. Here,
each inequality constraint appears in its own row. Now, we can represent the problem in the form
of an augmented matrix, which is called the initial simplex tableau.
Step 4: Identify the greatest negative entry in the bottom row, which helps to identify the pivot
column. The greatest negative entry in the bottom row defines the largest coefficient in the
objective function, which will help us to increase the value of the objective function as fastest as
possible.
Step 5: Compute the quotients. To calculate the quotient, we need to divide the entries in the far
right column by the entries in the first column, excluding the bottom row. The smallest quotient
identifies the row. The row identified in this step and the element identified in the step will be
taken as the pivot element.
Step 6: Carry out pivoting to make all other entries in the column is zero.
Step 7: If there are no negative entries in the bottom row, end the process. Otherwise, start from
step 4.
Step 8: Finally, determine the solution associated with the final simplex tableau.
Graphical method
The graphical method is used to optimize the two-variable linear programming. If the problem
has two decision variables, a graphical method is the best method to find the optimal solution. In
this method, the set of inequalities are subjected to constraints. Then the inequalities are plotted
in the XY plane. Once, all the inequalities are plotted in the XY graph, the intersecting region
will help to decide the feasible region. The feasible region will provide the optimal solution as
well as explains what all values our model can take.
The following steps are important to use this method.
Step 1: Formulate the LP (Linear programming) problem

Step 2: Construct a graph and plot the constraint lines

Step 3: Determine the valid side of each constraint line

Step 4: Identify the feasible solution region

Step 5: Plot the objective function on the graph

Step 6: Find the optimum point

#7. Discuss about the difference between Vogel’s Approximation method and Russel’s
Approximation method of finding the initial solution of a transportation problem.

 Russel’s Approximation Method (RAM):- Optimization begins by finding the highest

cost for each row and column in the transport table.
 RAM is an initial optimization method whose results are also better compared to the
North West Corner method and the least cost method.
 The Vogel Approximation Method is more optimum when compared with RAM.

 Vogel's Approximation method it gives the best initial Basic Feasible Solution (which
can be degenerate). All the other methods are either equal or inferior to Vogel's.
Moreover, degeneracy can happen in any method. Whenever any individual demand
quantity is equal to the individual supply quantity then degeneracy is going to take
place. BTW whatever one gets as the initial BFS, MODI will lead to Optimality.

 Vogel's Approximation Method tackles the problem of finding a good initial solution
by taking into account the costs associated with each alternative route, which is
something that Northwest Corner Rule did not do.Though VAM is not quite as simple
as Northwest Corner approach, but it facilitates a very good initial solution, one that
is often the optimal solution.

 VAM takes many more calculations to find an initial solution as compared to the
other methods, it almost produces a much better initial solution. 

 Hence VAM tends to minimize the total number of computations needed to reach an
optimal solution.

The completion steps Vogel's Approximation Method is as follows:

1. Arrange the value matrix of transport costs and the value of the capacity of each source into
columns and rows.
2. Find the two smallest cost of each row and column then calculate the difference of 2 the
smallest cost.
3. Finding 1 (one) largest difference value of the difference between row and column.
4. Selecting the cell with the lowest transportation cost that is in the column or row that has the
smallest value of the difference then allocate products/goods on the cell.
5. Repeat step 2 until step 4 until all products are distributed.
6. If the allocation has been completed then calculate the distribution cost.

 Russel’s Approximation Method (RAM): Solving the problem of transportation cost

optimization using Russel's approximation method can be said to have the same logic or
working method using the Vogels approximation method.
The steps of Russell's Approximation Method are described as follows:
1. Arrange the value matrix of transport costs and value of the capacity of each source into
columns and rows.
2. Find the highest cost value for each row and column
3. Subtract each cost value on the columns and rows at their highest cost.
4. Selecting the cell that has the greatest negative value from the calculation of step 2 (two) and
then allocates goods or products on the cell
5. Repeat step 2 until step 4 until all products are distributed.
6. If the allocation has been completed then calculate the distribution cost.

#8. A manufacturing company contemplates to produce two products, called X and Y, which can
be marketed at prevailing prices in any reasonable quantities without difficulty. It is known that
product X requires 10 and 5 man hours per unit in foundry and machine departments respectively
and that product Y requires 6 and 4 man hours per unit. The profit margin of X is Birr 30 per unit
and that of Y is Birr 20 per unit. It is estimated that there will be 1000 and 600 man hours
available in the foundry and machine departments respectively. Then,

A) Formulate the problem as linear programming problem model.

Solution: Objective function, Z=30x+20y,
Maximize Z
Constraints: x ≥ 0
y≥0
10x+6y ≤1000
5x+4y ≤ 600
B.How much of X and Y should be produced in order to maximize the company’s profit? (Use
graphic method method).

Corner points:
At origin(0,0)
Set x=0 and y=0
10(0)+6(0)+S1+0S2=1000
5(0)+4(0)+0S1+S2=600
Therefore, S1=1000
S2=600
Solution: x=0,y=0,s1=1000,s2=600
Profit:Z=30(0)+20(0)+0(1000)+600(0)=0

At (100,0)
Set x=100,y=0,s1=0
10x+6(0)+0+0s2=1000
5x+4(0)+0(0)+s2=600
Therefore,
10x=1000
5x+s2=600
Solution: x=100,y=0,s1=0,s2=100
Profit: Z=30(100)+20(0)+0(0)+0(100)=3000

At (40,100)
Set s1=0,s2=0
10x+6y=1000
5x+4y=600
Therefore, x=40,y=100
Profit:Z=30(40)+20(100)=3200

At(0,150)
Set x=0,s2=0
10(0)+6y+s1+0(0)=1000
5(0)+4y+0s1+0=600
Therefore,6y+s1=1000 and 4y=600
Solution:x=0,y=150, S1=850 and s2=0
Profit:30(0)+20(150)+0(850)+0=3000
At(0,500/3) and at(120,0) pints are infeasible. Because they violate the foundry and
machine constraints respectively.

Coordinates Z Values
(0,0) 0
(100,0) 3000
(40,100) 3200
(0,150) 3000

 The company should produce 40 products of X and 100 products of Y to obtain a benefit
of 3200 profits.
 B) Find the optimal solution using simplex method
Solution:
Max Z= 30X+20Y +0S1 +0S2
Subject to: 10X+6Y+ 0S1=1000
5X+4Y+ 0S2 = 600
X ≥ 0,Y ≥ 0,S1 ≥ 0 and S2 ≥ 0
For the initial solution simplex table
CBi Cj 30 20 0 0 solution Ratio
B.V. X Y S1 S2
0 S1 10 6 1 0 1000 0
0 S2 5 4 0 1 600 0
Zj 0 0 0 0 0
Cj-Zj 30 20 0 0

X1=0,X2=0,S1=1000,S2=600

Optimality Condition
For max:
all Cj -Zj ≤ 0

Iteration 1
Cj 30 20 0 0 Solution Ratio
CBi B.V. X Y S1 S2
0 S1 10 6 1 0 1000 1000/10=100
0 S2 5 4 0 1 600 600/5=120
 Zj 0 0 0 0 0 z=0
Cj-Zj 30 20 0 0 0
2
Zj = ∑ (CBi ) (aij) Cj-Zj=0
i=1
0×10 +0×5=0
0×6 + 0×4=0
0×1 +0×0=0
0×0 +0×1=0
0×1000+0×600=0

Iteration 2
CBi Cj 30 20 0 0 Solution Ratio
B.V. X Y S1 S2
30 X1 1 3/5 1/10 0 100 500/3
0 S2 0 1 -1/2 1 100 100
Zj 30 18 3 0 Z=3000
Cj-Zj 0 2 -3 0
Zj Cj-Zj =30-30=0,20-18=2,0-3= -3,0-0=0
30×1+0×0=30
30×3/5+0×1=18
30×1/10+0×-1/2=3
30×0++0×1=0
30×100+0×100=3000

Iteration 3
Zj=30×1+20×0=30 Cj-Zj=30-30=0,20-20=0,0-2= -2,0-2= -2
30×0+20×1=20
30×2/5+20×-1/2=2
30×-3/5+20×1=2
30×40+20×100=3200

CBi Cj 30 20 0 0 Solution Ratio

B.V. X Y S1 S2
30 X 1 0 2/5 -3/5 40 0
20 Y 0 1 -1/2 1 100 0
Zj 30 20 2 2 Z=3200
Cj-Zj 0 0 -2 -2

There is no any positive number greater than 0. so , in this case I started at the beginning by
taking Cj-Zj Therefore ,solution is : X=40, Y=100 and Z (Opt.)=3200

≠9. An oil company has three refineries that produce petrol which is then transported to Four
distribution centers. The availability of oil in barrels from these three refineries Is 16000, 20000
and 14000 respectively? The requirement of petrol in four centers is 10000, 10000, 12000 and
28000 respectively. The unit cost of transporting petrol From each refinery to each center is
given below
 W X Y Z
A 80 70 50 60
B 90 60 40 80
C 50 50 95 90
a) Find the initial basic feasible solution using VAM and RAM?
b) Suggest the transportation schedule that minimizes the total transportation cost using VAM?
c) What is the minimized total transportation cost using RAM?
Solution:
Total number of supply constraints are 3
Total number of demand constraints are 4

W X Y Z Supply
A 80 70 50 60 16000

B 90 60 40 80 20000

C 50 50 95 90 14000
Demand 10000 10000 12000 28000

Here Total Demand = 60000 is greater than Total Supply = 50000. So We add a dummy supply
constraint with 0 unit cost and with allocation 10000.

Now, The modified table is

W X Y Z Supply

A 80 70 50 60 16000

B 90 60 40 80 20000

C 50 50 95 90 14000

Ddummy 0 0 0 0 10000

Deman 1000 1000 1200 2800

d 0 0 0 0
Table-1
W X Y Z Supply Row Penalty

A 80 70 50 60 16000 10=60-50

B 90 60 40 80 20000 20=60-40

C 50 50 95 90 14000 0=50-50

Ddummy 0 0 0 0 10000 0=0-0

Demand 10000 10000 12000 28000

Column 50=50- 40=40-

50=50-0 60=60-0
Penalty 0 0

The maximum penalty, 60, occurs in column Z.

The minimum cij in this column is c44=0.
The maximum allocation in this cell is min(10000,28000) = 10000.
It satisfy supply of Ddummy and adjust the demand of Z from 28000 to 18000 (28000 - 10000=18000).

Table-2
W X Y Z Supply Row Penalty

A 80 70 50 60 16000 10=60-50

B 90 60 40 80 20000 20=60-40

C 50 50 95 90 14000 0=50-50

Ddummy 0 0 0 0(10000) 0 --

Demand 10000 10000 12000 18000

Column 20=80-
30=80-50 10=60-50 10=50-40
Penalty 60

The maximum penalty, 30, occurs in column W.

The minimum cij in this column is c31=50.
The maximum allocation in this cell is min(14000,10000) = 10000.
It satisfy demand of W and adjust the supply of C  from 14000 to 4000 (14000 - 10000=4000).
Table-3
W X Y Z Supply Row Penalty

A 80 70 50 60 16000 10=60-50

B 90 60 40 80 20000 20=60-40

C 50(10000) 50 95 90 4000 40=90-50

Ddummy 0 0 0 0(10000) 0 --

Demand 0 10000 12000 18000

Column 10=60- 20=80-

-- 10=50-40
Penalty 50 60

The maximum penalty, 40, occurs in row C.

The minimum cij in this row is c32=50.
The maximum allocation in this cell is min(4000,10000) = 4000.
It satisfy supply of C and adjust the demand of B from 10000 to 6000 (10000 - 4000=6000).

Table-4
W X Y Z Supply Row Penalty

A 80 70 50 60 16000 10=60-50

B 90 60 40 80 20000 20=60-40

C 50(10000) 50(4000) 95 90 0 --

Ddummy 0 0 0 0(10000) 0 --

Demand 0 6000 12000 18000

Column 10=70- 20=80-

-- 10=50-40
Penalty 60 60

The maximum penalty, 20, occurs in row B.

The minimum cij in this row is c23=40.
The maximum allocation in this cell is min(20000,12000) = 12000.
It satisfy demand of Y and adjust the supply of B  from 20000 to 8000 (20000 - 12000=8000).
Table-5
W X Y Z Supply Row Penalty

A 80 70 50 60 16000 10=70-60

40(12000
B 90 60 80 8000 20=80-60
)

50(10000
C 50(4000) 95 90 0 --
)

Ddummy 0 0 0 0(10000) 0 --

Demand 0 6000 0 18000

Column
-- 10=70-60 -- 20=80-60
Penalty

The maximum penalty, 20, occurs in column Z.

The minimum cij in this column is c14=60.
The maximum allocation in this cell is min(16000,18000) = 16000.
It satisfy supply of A  and adjust the demand of Z  from 18000 to 2000 (18000 - 16000=2000).

Table-6
W X Y Z Supply Row Penalty

60(16000
A 80 70 50 0 --
)

40(12000
B 90 60 80 8000 20=80-60
)

C 50(10000) 50(4000) 95 90 0 --

Ddummy 0 0 0 0(10000) 0 --

Demand 0 6000 0 2000

Column
-- 60 -- 80
Penalty

The maximum penalty, 80, occurs in column Z.

The minimum cij in this column is c24=80.

The maximum allocation in this cell is min(8000,2000) = 2000.
It satisfy demand of Z  and adjust the supply of B  from 8000 to 6000 (8000 - 2000=6000).
Table -7
W X Y Z Supply Row Penalty

60(16000
A 80 70 50 0 --
)

40(12000
B 90 60 80(2000) 6000 60
)

C 50(10000) 50(4000) 95 90 0 --

Sdummy 0 0 0 0(10000) 0 --

Demand 0 6000 0 0

Column
-- 60 -- --
Penalty

The maximum penalty, 60, occurs in row B.

The minimum cij in this row is c22=60.
The maximum allocation in this cell is min(6000,6000) = 6000.
It satisfy supply of B  and demand of X.

Initial feasible solution is

W X Y Z Supply Row Penalty

60(16000
A 80 70 50 16000 10 | 10 | 10 | 10 | 10 | -- | -- |
)

B 90 60(6000) 40(12000) 80(2000) 20000 20 | 20 | 20 | 20 | 20 | 20 | 60 |

50(10000
C 50(4000) 95 90 14000  0 |  0 | 40 | -- | -- | -- | -- |
)

Ddummy 0 0 0 0(10000) 10000  0 | -- | -- | -- | -- | -- | -- |

Demand 10000 10000 12000 28000

Column 50 50 40 60
Penalty 30 10 10 20
-- 10 10 20
-- 10 10 20
 -- 10 -- 20
-- 60 -- 80
-- 60 -- --

The minimum total transportation

cost =60×16000+60×6000+40×12000+80×2000+50×10000+50×4000+0×10000=2660000

Here, the number of allocated cells = 7 is equal to m + n - 1 = 4 + 4 - 1 = 7

This solution is non-degenerate

Optimal solution calculation:

Optimality test using modi method...
Allocation Table is

W X Y Z Supply

A 80 70 50 60 (16000) 16000

40 (12000
B 90 60 (6000) 80 (2000) 20000
)

50 (10000
C 50 (4000) 95 90 14000
)

Ddummy 0 0 0 0 (10000) 10000

Demand 10000 10000 12000 28000

Iteration-1 of optimality test

1. Find ui and vj for all occupied cells(i,j), where cij=ui+vj
1. Substituting, u2=0, we get

2.c22=u2+v2⇒v2=c22-u2⇒v2=60-0⇒v2=60

3.c32=u3+v2⇒u3=c32-v2⇒u3=50-60⇒u3=-10

4.c31=u3+v1⇒v1=c31-u3⇒v1=50+10⇒v1=60

 5.c23=u2+v3⇒v3=c23-u2⇒v3=40-0⇒v3=40

6.c24=u2+v4⇒v4=c24-u2⇒v4=80-0⇒v4=80

7.c14=u1+v4⇒u1=c14-v4⇒u1=60-80⇒u1=-20

8.c44=u4+v4⇒u4=c44-v4⇒u4=0-80⇒u4=-80

W X Y Z Supply ui
 A 80 70 50 60 (16000) 16000 u1=-20

B 90 60 (6000) 40 (12000) 80 (2000) 20000 u2=0

50 (10000
C 50 (4000) 95 90 14000 u3=-10
)

Ddummy 0 0 0 0 (10000) 10000 u4=-80

Demand 10000 10000 12000 28000

vj v1=60 v2=60 v3=40 v4=80

( )
. Find dij for all unoccupied cells(i,j), where dij=cij- ui+vj

( )
1.d11=c11- u1+v1 =80-(-20+60)=40

( )
2.d12=c12- u1+v2 =70-(-20+60)=30

( )
3.d13=c13- u1+v3 =50-(-20+40)=30

( )
4.d21=c21- u2+v1 =90-(0+60)=30

( )
5.d33=c33- u3+v3 =95-(-10+40)=65

( )
6.d34=c34- u3+v4 =90-(-10+80)=20

( )
7.d41=c41- u4+v1 =0-(-80+60)=20

( )
8.d42=c42- u4+v2 =0-(-80+60)=20

( )
9.d43=c43- u4+v3 =0-(-80+40)=40

W X Y Z Supply ui

A 80 [40] 70 [30] 50 [30] 60 (16000) 16000 u1=-20

B 90 [30] 60 (6000) 40 (12000) 80 (2000) 20000 u2=0

 50 (10000
C 50 (4000) 95 [65] 90 [20] 14000 u3=-10
)

Ddummy 0 [20] 0 [20] 0 [40] 0 (10000) 10000 u4=-80

Demand 10000 10000 12000 28000

vj v1=60 v2=60 v3=40 v4=80

Since all dij ≥ 0.

So, final optimal solution is arrived.

W X Y Z Supply

A 80 70 50 60 (16000) 16000

B 90 60 (6000) 40 (12000) 80 (2000) 20000

C 50 50 (4000) 95 90 14000

dummy 0 0 0 0 (10000) 10000

Demand 10000 10000 12000 28000

The minimum total transportation
cost =60×16000+60×6000+40×12000+80×2000+50×10000+50×4000+0×10000=2660000.

The initial basic solution and the minimized total transportation cost using RAM

Find Solution using Russell's Approximation method

D1 D2 D3 D4 Supply
S1 80 70 50 60 16000
S2 90 60 40 80 20000
S3 50 50 95 90 14000
Demand 10000 10000 12000 28000

Solution:
TOTAL number of supply constraints : 3
TOTAL number of demand constraints : 4
Problem Table is
 D1 D2 D3 D4 Supply

S1 80 70 50 60 16000

S2 90 60 40 80 20000

S3 50 50 95 90 14000

Demand 10000 10000 12000 28000

Here Total Demand = 60000 is greater than Total Supply = 50000. So We add a dummy supply
constraint with 0 unit cost and with allocation 10000.
Now, The modified table is
D1 D2 D3 D4 Supply

S1 80 70 50 60 16000

S2 90 60 40 80 20000

S3 50 50 95 90 14000

Sdummy 0 0 0 0 10000

Demand 10000 10000 12000 28000

Table-1: Calculate ˉUi and ˉVj (where ˉUi is the largest cost in row and ˉVj is the largest cost in
column)

D1 D2 D3 D4 Supply ˉU
i

S1 80 70 50 60 16000 80

S2 90 60 40 80 20000 90

S3 50 50 95 90 14000 95

Sdummy 0 0 0 0 10000 0

1200
Demand 10000 10000 28000
0

ˉVj 90 70 95 90

(
2. Compute reduced cost of each cell Δij, where Δij=cij- ˉUi+ˉVj )
 ( )
1.Δ11=c11- ˉU1+ˉV1 =80-(80+90)=-90

( )
2.Δ12=c12- ˉU1+ˉV2 =70-(80+70)=-80

( )
3.Δ13=c13- ˉU1+ˉV3 =50-(80+95)=-125

( )
4.Δ14=c14- ˉU1+ˉV4 =60-(80+90)=-110

( )
5.Δ21=c21- ˉU2+ˉV1 =90-(90+90)=-90

( )
6.Δ22=c22- ˉU2+ˉV2 =60-(90+70)=-100

( )
7.Δ23=c23- ˉU2+ˉV3 =40-(90+95)=-145

( )
8.Δ24=c24- ˉU2+ˉV4 =80-(90+90)=-100

( )
9.Δ31=c31- ˉU3+ˉV1 =50-(95+90)=-135

( )
10.Δ32=c32- ˉU3+ˉV2 =50-(95+70)=-115

( )
11.Δ33=c33- ˉU3+ˉV3 =95-(95+95)=-95

( )
12.Δ34=c34- ˉU3+ˉV4 =90-(95+90)=-95

( )
13.Δ41=c41- ˉU4+ˉV1 =0-(0+90)=-90

( )
14.Δ42=c42- ˉU4+ˉV2 =0-(0+70)=-70

( )
15.Δ43=c43- ˉU4+ˉV3 =0-(0+95)=-95

( )
16.Δ44=c44- ˉU4+ˉV4 =0-(0+90)=-90

D1 D2 D3 D4 Supply ˉU
 i

S1 80 [-90] 70 [-80] 50 [-125] 60 [-110] 16000 80

60 [-
S2 90 [-90] 40 [-145] 80 [-100] 20000 90
100]

50 [-
S3 50 [-135] 95 [-95] 90 [-95] 14000 95
115]

Sdummy 0 [-90] 0 [-70] 0 [-95] 0 [-90] 10000 0

Demand 10000 10000 12000 28000

ˉVj 90 70 95 90

The most negative Δij is -145 in cell S2D3

The allocation to this cell is min(20000,12000) = 12000.

This satisfies the entire demand of D3 and leaves 20000 - 12000=8000 units with S2

Table-1: This leads to the following table

D1 D2 D3 D4 Supply

S1 80 70 50 60 16000

S2 90 60 40 (12000) 80 8000

S3 50 50 95 90 14000

Sdummy 0 0 0 0 10000

1000
Demand 10000 0 28000
0

Table-2: Calculate ˉUi and ˉVj (where ˉUi is the largest cost in row and ˉVj is the largest cost in
column)

D1 D2 D3 D4 Supply ˉU
i
 S1 80 70 50 60 16000 80

S2 90 60 40(12000) 80 8000 90

S3 50 50 95 90 14000 90

Sdummy 0 0 0 0 10000 0

1000 2800
Demand 10000 0
0 0

ˉVj 90 70 -- 90

(
2. Compute reduced cost of each cell Δij, where Δij=cij- ˉUi+ˉVj )

( )
1.Δ11=c11- ˉU1+ˉV1 =80-(80+90)=-90

( )
2.Δ12=c12- ˉU1+ˉV2 =70-(80+70)=-80

( )
3.Δ14=c14- ˉU1+ˉV4 =60-(80+90)=-110

( )
4.Δ21=c21- ˉU2+ˉV1 =90-(90+90)=-90

( )
5.Δ22=c22- ˉU2+ˉV2 =60-(90+70)=-100

( )
6.Δ24=c24- ˉU2+ˉV4 =80-(90+90)=-100

( )
7.Δ31=c31- ˉU3+ˉV1 =50-(90+90)=-130

( )
8.Δ32=c32- ˉU3+ˉV2 =50-(90+70)=-110

( )
9.Δ34=c34- ˉU3+ˉV4 =90-(90+90)=-90

( )
10.Δ41=c41- ˉU4+ˉV1 =0-(0+90)=-90
 ( )
11.Δ42=c42- ˉU4+ˉV2 =0-(0+70)=-70

( )
12.Δ44=c44- ˉU4+ˉV4 =0-(0+90)=-90

Suppl ˉU
D1 D2 D3 D4
y i

S1 80 [-90] 70 [-80] 50 60 [-110] 16000 80

40(12000
S2 90 [-90] 60 [-100] 80 [-100] 8000 90
)

50 [-
S3 50 [-110] 95 90 [-90] 14000 90
130]

Sdummy 0 [-90] 0 [-70] 0 0 [-90] 10000 0

Demand 10000 10000 0 28000

ˉVj 90 70 -- 90

The most negative Δij is -130 in cell S3D1

The allocation to this cell is min(14000,10000) = 10000.

This satisfies the entire demand of D1 and leaves 14000 - 10000=4000 units with S3

Table-2: This leads to the following table

D1 D2 D3 D4 Supply

S1 80 70 50 60 16000

S2 90 60 40 (12000) 80 8000

S3 50 (10000) 50 95 90 4000

Sdummy 0 0 0 0 10000

1000
Demand 0 0 28000
0

Table-3: Calculate ˉUi and ˉVj (where ˉUi is the largest cost in row and ˉVj is the largest cost in
column)

D1 D2 D3 D4 Supply ˉU
i
 S1 80 70 50 60 16000 70

S2 90 60 40(12000) 80 8000 80

50(10000
S3 50 95 90 4000 90
)

Sdummy 0 0 0 0 10000 0

2800
Demand 0 10000 0
0

ˉVj -- 70 -- 90

(
2. Compute reduced cost of each cell Δij, where Δij=cij- ˉUi+ˉVj )
( )
1.Δ12=c12- ˉU1+ˉV2 =70-(70+70)=-70

( )
2.Δ14=c14- ˉU1+ˉV4 =60-(70+90)=-100

( )
3.Δ22=c22- ˉU2+ˉV2 =60-(80+70)=-90

( )
4.Δ24=c24- ˉU2+ˉV4 =80-(80+90)=-90

( )
5.Δ32=c32- ˉU3+ˉV2 =50-(90+70)=-110

6.Δ34=c34-(ˉU3+ˉV4)=90-(90+90)=-90

( )
7.Δ42=c42- ˉU4+ˉV2 =0-(0+70)=-70

( )
8.Δ44=c44- ˉU4+ˉV4 =0-(0+90)=-90

D1 D2 D3 D4 Supply ˉU
i

S1 80 70 [-70] 50 60 [-100] 16000 70

40(12000
S2 90 60 [-90] 80 [-90] 8000 80
)

S3 50(10000 50 [-110] 95 90 [-90] 4000 90

 )

Sdummy 0 0 [-70] 0 0 [-90] 10000 0

Demand 0 10000 0 28000

ˉVj -- 70 -- 90

The most negative Δij is -110 in cell S3D2

The allocation to this cell is min(4000,10000) = 4000.

This exhausts the capacity of S3 and leaves 10000 - 4000=6000 units with D2

Table-3: This leads to the following table

D1 D2 D3 D4 Supply

S1 80 70 50 60 16000

40 (12000
S2 90 60 80 8000
)

S3 50 (10000) 50 (4000) 95 90 0

Sdummy 0 0 0 0 10000

Demand 0 6000 0 28000

Table-4: Calculate ˉUi and ˉVj (where ˉUi is the largest cost in row and ˉVj is the largest cost in
column)

D1 D2 D3 D4 Supply ˉU
i

S1 80 70 50 60 16000 70

40(12000
S2 90 60 80 8000 80
)

50(10000
S3 50(4000) 95 90 0 --
)

Sdummy 0 0 0 0 10000 0

Demand 0 6000 0 28000

ˉVj -- 70 -- 80
 (
2. Compute reduced cost of each cell Δij, where Δij=cij- ˉUi+ˉVj )
( )
1.Δ12=c12- ˉU1+ˉV2 =70-(70+70)=-70

( )
2.Δ14=c14- ˉU1+ˉV4 =60-(70+80)=-90

( )
3.Δ22=c22- ˉU2+ˉV2 =60-(80+70)=-90

( )
4.Δ24=c24- ˉU2+ˉV4 =80-(80+80)=-80

( )
5.Δ42=c42- ˉU4+ˉV2 =0-(0+70)=-70

( )
6.Δ44=c44- ˉU4+ˉV4 =0-(0+80)=-80

D1 D2 D3 D4 Supply ˉU
i

S1 80 70 [-70] 50 60 [-90] 16000 70

40(12000
S2 90 60 [-90] 80 [-80] 8000 80
)

50(10000
S3 50(4000) 95 90 0 --
)

Sdummy 0 0 [-70] 0 0 [-80] 10000 0

Demand 0 6000 0 28000

ˉVj -- 70 -- 80

The most negative Δij is -90 in cell S1D4

The allocation to this cell is min(16000,28000) = 16000.

This exhausts the capacity of S1 and leaves 28000 - 16000=12000 units with D4

 Table-4: This leads to the following table

D1 D2 D3 D4 Supply

S1 80 70 50 60 (16000) 0
 40 (12000
S2 90 60 80 8000
)

50 (10000
S3 50 (4000) 95 90 0
)

Sdummy 0 0 0 0 10000

Demand 0 6000 0 12000

Table-5: Calculate ˉUi and ˉVj (where ˉUi is the largest cost in row and ˉVj is the largest cost in
column)

D1 D2 D3 D4 Supply ˉU
i

60(16000
S1 80 70 50 0 --
)

S2 90 60 40(12000) 80 8000 80

50(10000
S3 50(4000) 95 90 0 --
)

Sdummy 0 0 0 0 10000 0

Demand 0 6000 0 12000

ˉVj -- 60 -- 80

2. Compute reduced cost of each cell Δij, where Δij=cij- ˉUi+ˉVj( )

( )
1.Δ22=c22- ˉU2+ˉV2 =60-(80+60)=-80

( )
2.Δ24=c24- ˉU2+ˉV4 =80-(80+80)=-80

( )
3.Δ42=c42- ˉU4+ˉV2 =0-(0+60)=-60

4.Δ44=c44-(ˉU4+ˉV4)=0-(0+80)=-80

D1 D2 D3 D4 Supply ˉU
i
 60(16000
S1 80 70 50 0 --
)

S2 90 60 [-80] 40(12000) 80 [-80] 8000 80

50(10000
S3 50(4000) 95 90 0 --
)

Sdummy 0 0 [-60] 0 0 [-80] 10000 0

Demand 0 6000 0 12000

ˉVj -- 60 -- 80

The most negative Δij is -80 in cell SdummyD4

The allocation to this cell is min(10000,12000) = 10000.

This exhausts the capacity of Sdummy and leaves 12000 - 10000=2000 units with D4

Table-5: This leads to the following table

D1 D2 D3 D4 Supply

S1 80 70 50 60 (16000) 0

40 (12000
S2 90 60 80 8000
)

50 (10000
S3 50 (4000) 95 90 0
)

Sdummy 0 0 0 0 (10000) 0

Demand 0 6000 0 2000

Table-6: Calculate ˉUi and ˉVj (where ˉUi is the largest cost in row and ˉVj is the largest cost in
column)

D1 D2 D3 D4 Supply ˉU
i

60(16000
S1 80 70 50 0 --
)

S2 90 60 40(12000) 80 8000 80

S3 50(10000 50(4000) 95 90 0 --
 )

Sdummy 0 0 0 0(10000) 0 --

Demand 0 6000 0 2000

ˉVj -- 60 -- 80

(
2. Compute reduced cost of each cell Δij, where Δij=cij- ˉUi+ˉVj )
( )
1.Δ22=c22- ˉU2+ˉV2 =60-(80+60)=-80

( )
2.Δ24=c24- ˉU2+ˉV4 =80-(80+80)=-80

D1 D2 D3 D4 Supply ˉU
i

60(16000
S1 80 70 50 0 --
)

S2 90 60 [-80] 40(12000) 80 [-80] 8000 80

50(10000
S3 50(4000) 95 90 0 --
)

Sdummy 0 0 0 0(10000) 0 --

Demand 0 6000 0 2000

ˉVj -- 60 -- 80

The most negative Δij is -80 in cell S2D2

The allocation to this cell is min(8000,6000) = 6000.

This satisfies the entire demand of D2 and leaves 8000 - 6000=2000 units with S2
Table-6: This leads to the following table
D1 D2 D3 D4 Supply

S1 80 70 50 60 (16000) 0

40 (12000
S2 90 60 (6000) 80 2000
)
 50 (10000
S3 50 (4000) 95 90 0
)

Sdummy 0 0 0 0 (10000) 0

Demand 0 0 0 2000

Table-7: Calculate ˉUi and ˉVj (where ˉUi is the largest cost in row and ˉVj is the largest cost in
column)

D1 D2 D3 D4 Supply ˉU
i

60(16000
S1 80 70 50 0 --
)

S2 90 60(6000) 40(12000) 80 2000 80

50(10000
S3 50(4000) 95 90 0 --
)

Sdummy 0 0 0 0(10000) 0 --

Demand 0 0 0 2000

ˉVj -- -- -- 80

(
2. Compute reduced cost of each cell Δij, where Δij=cij- ˉUi+ˉVj )
( )
1.Δ24=c24- ˉU2+ˉV4 =80-(80+80)=-80

D1 D2 D3 D4 Supply ˉU
i

60(16000
S1 80 70 50 0 --
)

S2 90 60(6000) 40(12000) 80 [-80] 2000 80

50(10000
S3 50(4000) 95 90 0 --
)

Sdummy 0 0 0 0(10000) 0 --

Demand 0 0 0 2000

ˉVj -- -- -- 80
The most negative Δij is -80 in cell S2D4

The allocation to this cell is min(2000,2000) = 2000.

Table-7: This leads to the following table
D1 D2 D3 D4 Supply

S1 80 70 50 60 (16000) 0

40 (12000
S2 90 60 (6000) 80 (2000) 0
)

50 (10000
S3 50 (4000) 95 90 0
)

Sdummy 0 0 0 0 (10000) 0

Demand 0 0 0 0

Initial feasible solution is

D1 D2 D3 D4 Supply

S1 80 70 50 60 (16000) 16000

40 (12000
S2 90 60 (6000) 80 (2000) 20000
)

50 (10000
S3 50 (4000) 95 90 14000
)

Sdummy 0 0 0 0 (10000) 10000

Demand 10000 10000 12000 28000

The minimum total transportation

cost =60×16000+60×6000+40×12000+80×2000+50×10000+50×4000+0×10000=2660000

Here, the number of allocated cells = 7 is equal to m + n - 1 = 4 + 4 - 1 = 7

∴ This solution is non-degenerate