ASSIGNMENTS

MB 0032
(3 credits)
Set 1
Marks 60
OPERATIONS RESEARCH
Note: Each Question carries 10 marks
1. Describe the broad classification of Operations Research models in detai
ls. Name the different steps needed in OR approach of problem solving?
Broad Classification of Operations Research models are:-
Linear Programming Models
Network Flow Programming Models
Integer Programming Models
Nonlinear Programming Models
Probability Models
Discrete Time Markov Chain Models
Continuous Time Markov Chain Models
Queuing Models
Simulation Models
Most operations research studies involve the construction of a mathematical mode
l. The model is a collection of logical and mathematical relationships that repr
esents aspects of the situation under study. Models describe important relations
hips between variables, include an objective function with which alternative sol
utions are evaluated, and constraints that restrict solutions to feasible values
.
Although the analyst would hope to study the broad implications of the problem u
sing a systems approach, a model cannot include every aspect of a situation. A m
odel is always an abstraction that is of necessity simpler than the real situati
on. Elements that are irrelevant or unimportant to the problem are to be ignored
, hopefully leaving sufficient detail so that the solution obtained with the mod
el has value with regard to the original problem.
Models must be both tractable, capable of being solved, and valid, representativ
e of the original situation. These dual goals are often contradictory and are no
t always attainable. It is generally true that the most powerful solution method
s can be applied to the simplest, or most abstract, model.
We provide in this section, a description of the various types of models used by
operations research analysts. The division is based on the mathematical form of
the model. All the models described here are solved with Excel add-ins describe
d in the Computation section of this site. In some cases, the methods used to so
lve a model are described in the Methods section. Student exercises for creating
models are in the Problems section. Additional models related to problems arisi
ng in Operations Management and Industrial Engineering are in the OM/IE section.
2. Describe the Penalty Cost method (Big M Method) for solving L.P.P.
The Big M Method :
If an LP has any > or = constraints, a starting basic feasible solution may not
be readily apparent.
The Big M method is a version of the Simplex Algorithm that first finds a basic
feasible solution by adding "artificial" variables to the problem. The objective
function of the original LP must, of course, be modified to ensure that the art
ificial variables are all equal to 0 at the conclusion of the simplex algorithm.
Steps
1. Modify the constraints so that the RHS of each constraint is nonnegative (Thi
s requires that each constraint with a negative RHS be multiplied by -1. Remembe
r that if you multiply an inequality by any negative number, the direction of th
e inequality is reversed!). After modification, identify each constraint as a <,
>, or = constraint. 2. Convert each inequality constraint to standard form (If
constraint i is a < constraint, we add aslack variable si; and if constraint i i
s a > constraint, we subtract an excess variable ei).
3. Add an artificial variable ai to the constraints identified as > or = constra
ints at the end of
Step 1. Also add the sign restriction ai > 0.
4. If the LP is a max problem, add (for each artificial variable) -Mai to the ob
jective function where M denote a very large positive number.
5. If the LP is a min problem, add (for each artificial variable) Mai to the obj
ective function.
6. Solve the transformed problem by the simplex . Since each artificial variable
will be in the starting basis, all artificial variables must be eliminated from
row 0 before beginning the simplex. Now (In choosing the entering variable, rem
ember that M is a very large positive number!).
If all artificial variables are equal to zero in the optimal solution, we have f
ound the optimal solution to the original problem.
If any artificial variables are positive in the optimal solution, the original p
roblem is infeasible

3. Why Duality concept is important in OR? Describe the economic importance of t

he Duality concept.
THE DUALITY CONCEPT in OR:
Accounting is basically the entire sequence of processes of depicting and record
ing the economic activities of a business enterprise, preparing the financial st
atements, and communicating them to users. Since the 1960s, however, there has b
een a tendency to emphasize only one phase of the sequence of processes that com
prise accounting, namely the relationship between the financial statements and t
he user, and to regard the remaining processes as subsidiary to that end. In thi
s paper I shall argue that the duality which governs the process of depiction of
economic activities is a fundamental characteristic of accounting, and I shall
address methodological issues concerning the basic equations, which represent du
ality concepts concretely. The basic equation has been formulated in various way
s, but essentially we may discern two viewpointsrone, which admits of only one f
ormulation, and a second, which hypothetically allows for multiple formulations
whose explanatory power is validated through practice. I shall argue here, in th
e first place, that the former viewpoint is inadequate both for calculative obje
ctives and in its methodology of theory construction, and go on to demonstrate t
he validity of the latter viewpoint. Secondly, I shall consider four theories, t
he capital equation, Walb theory, the balance sheet equation and the business ca
pital equation(or trial balance equation), and discuss their explanatory power b
y reference to the construct of calculative objects. In conclusion, I shall argu
e that the business capital equation has the highest explanatory power.
Accounting is basically the entire sequence of processes of depicting and record
ing the economic activities of a business enterprise, preparing the financial st
atements, and communicating them to users. Since the 1960s, however, there has b
een a tendency to emphasize only one phase of the sequence of processes that com
prise accounting, namely the relationship between the financial statements and t
he user, and to regard the remaining processes as subsidiary to that end. In thi
s paper I shall argue that the duality which governs the process of depiction of
economic activities is a fundamental characteristic of accounting, and I shall
address methodological issues concerning the basic equations, which represent du
ality concepts concretely. The basic equation has been formulated in various way
s, but essentially we may discern two viewpointsrone, which admits of only one f
ormulation, and a second, which hypothetically allows for multiple formulations
whose explanatory power is validated through practice. I shall argue here, in th
e first place, that the former viewpoint is inadequate both for calculative obje
ctives and in its methodology of theory construction, and go on to demonstrate t
he validity of the latter viewpoint. Secondly, I shall consider four theories, t
he capital equation, Walb theory, the balance sheet equation and the business ca
pital equation(or trial balance equation), and discuss their explanatory power b
y reference to the construct of calculative objects. In conclusion, I shall argu
e that the business capital equation has the highest explanatory power.
4. Describe the Matrix Minimum method of finding the initial basic feasible solu
tion in the transportation problem.
Matrix Minimum Method
Matrix minimum (Least cost) method is a method for computing a basic feasible so
lution of a transportation problem, where the basic variables are chosen accordi
ng to the unit cost of transportation. This method is very useful because it red
uces the computation and the time required to determine the optimal solution. Th
e following steps summarize the approach.
Steps
Identify the box having minimum unit transportation cost (cij).
If the minimum cost is not unique, then you are at liberty to choose any cell.
Choose the value of the corresponding xij as much as possible subject to the cap
acity and requirement constraints. Repeat steps 1-3 until all restrictions are s
atisfied.
Example
Consider the transportation problem presented in the following table:

Solution.
We observe that c21 =2, which is the minimum transportation cost. So x21 = 20.
The demand for the first column is satisfied. The allocation is shown in the fol
lowing table.
Table 1

Now we observe that c24 =2, which is the minimum transportation cost, so x24 = 5
5.
The supply for the second row is exhausted.

Table 2

Proceeding in this way, we observe that x34 = 5, x12 = 20, x13 = 30, x33 = 20. T
he resulting feasible solution is shown in the following table.
Final Table
Number of basic variables = m + n –1 = 3 + 4 – 1 = 6.
Initial basic feasible solution
The total transportation cost associated with this solution is calculated as giv
en below:
20 X 2 + 20 X 5 + 30 X 7 + 55 X 2 + 20 X 9 + 5 X 2 = 650.

5. What do you understand by the Integer Programming Problem? Describe the Gomor
y’s All-I.P.P. method for solving the I.P.P. problem.
When formulating LP s we often found that, strictly, certain variables should ha
ve been regarded as taking integer values but, for the sake of convenience, we l
et them take fractional values reasoning that the variables were likely to be so
large that any fractional part could be neglected. Whilst this is acceptable in
some situations, in many cases it is not, and in such cases we must find a nume
ric solution in which the variables take integer values. Problems in which this
is the case are called integer programs (IP s) and the subject of solving such
programs is called integer programming (also referred to by the initials IP). I
P s occur frequently because many decisions are essentially discrete (such as ye
s/no, go/no-go) in that one (or more) options must be chosen from a finite set o
f alternatives. Note here that problems in which some variables can take only in
teger values and some variables can take fractional values are called mixed-inte
ger programs (MIP s). As for formulating LP s the key to formulating IP s is pra
ctice. Although there are a number of standard "tricks" available to cope with s
ituations that often arise in formulating IP s it is probably true to say that f
ormulating IP s is a much harder task than formulating LP s.
Gomory s cutting plane method for integer programming:
The idea behind Gomory s method is to initially neglect the integrality requirem
ents and solve the corresponding linear programming problem, for instance with t
he simplex method. This will give an optimal vector bx. If all coefficients of b
x happen to be integer, it is also optimal for the integer programming problem,
which thereby is solved. What if this is not the case? An obvious suggestion wou
ld be to try to round bx to an integer solution. That, however, is too naive. Th
e rounded solution might not even be feasible, and even if it is, the optimal in
teger solution is typically not among these rounded vectors. Moreover, the knowl
edge of those rounded solutions that are feasible is too limited to facilitate e
fficient selection of an optimal one. Still, rounding is not such a bad idea, th
at is, as long as we do not round bx. We will use rounding to and a cutting plan
e. A cutting plane is a new inequality that is satisfied by all integral solutio
ns of the original system, but is violated by bx. Gomory showed that such a cutt
ing plane can always be found as follows. Suppose that bx is a vertex of the set
of all solutions to the given system of inequalities. The simplex method does g
uarantee that that is the case and that bx has some nonintegral coeficients. The
n there exists an inequality am+1x b that is satisfied by all solutions to the s
ystem (also the nonintegral ones) such that the following holds: am+1bx = b, all
coeficients of am+1 are integral, and b is not an integer. Gomory also showed t
hat the output of the simplex method produces such an inequality. Next, round b
down to the nearest integer, bm+1. The resulting inequality am+1x fi bm+1 is the
desired cutting plane. Gomory s cutting plane method for integer programming ad
ds this cutting plane to the system and iterates the whole procedure. Gomory sho
wed that alternately applying the simplex method and adding cutting planes even
tually leads to a system for which the simplex method will give an integer optim
um. Until this day basically all software for either general or specific integer
programming problems is based on cutting plane methods2|not necessarily Gomory
s cutting planes, but de nitely inspired by them. The application of Gomory s id
ea to specific problem classes requires good understanding of the applications
and the methods; genuine engineering for mathematicians. Besides algorithmic con
sequences, Gomory s cutting plane method also led to a better understanding of t
he geometric aspects of integer solutions to systems of inequalities. These insi
ghts still inspire mathematicians working on integer programming and polyhedral
zombinatorics (which is the application of linear algebra to combinatorial probl
ems).
ASSIGNMENTS
MB 0032
(3 credits)
Set 2
Marks 60
OPERATIONS RESEARCH
Note: Each Question carries 10 marks
1. Describe in details the OR approach of problem solving. What are the limitati
ons of the Operations Research?
Operations Research (OR) in North America, South Africa and Australia, and Opera
tional Research in Europe, is an interdisciplinary branch of applied mathematics
and formal science that uses methods such as mathematical modeling, statistics,
and algorithms to arrive at optimal or near optimal solutions to complex proble
ms. It is typically concerned with optimizing the maxima (profit, assembly line
performance, crop yield, bandwidth, etc) or minima (loss, risk, etc.) of some ob
jective function. Operations research helps management achieve its goals using s
cientific methods.
Scope of operations research
• critical path analysis or project planning: identifying those processes
in a complex project which affect the overall duration of the project
• designing the layout of a factory for efficient flow of materials
• constructing a telecommunications network at low cost while still guaran
teeing QoS (quality of service) or QoE (Quality of Experience) if particular con
nections become very busy or get damaged
• road traffic management and one way street allocations i.e. allocation
problems.
• determining the routes of school buses (or city buses) so that as few bu
ses are needed as possible
• designing the layout of a computer chip to reduce manufacturing time (th
erefore reducing cost)
• managing the flow of raw materials and products in a supply chain based
on uncertain demand for the finished products
• efficient messaging and customer response tactics
• robotizing or automating human-driven operations processes
• globalizing operations processes in order to take advantage of cheaper m
aterials, labor, land or other productivity inputs
• managing freight transportation and delivery systems scheduling:
o personnel staffing
o manufacturing steps
o project tasks
• blending of raw materials in oil refineries
2. What are the characteristics of the standard form of L.P.P.? What is the
standard form of L.P.P.? State the fundamental theorem of L.P.P.
The Standard Linear Programming Problem
Actually, the problem given above has some special and well known features like
a linear cost function f (x) to be minimized
and the constraints are also linear inequalities
This problem can be expressed in matrix form as
Minimize
Subject to
Such programs are called linear programming problems and they hold a very large
place in daily applications of mathematical programming.
In practice, all known methods for solving linear programming problems have been
developed for
Min
Subject to
and this is the standard linear programming (L.P.) problem. An inequality constr
ained problem can be converted into the standard form by
Minimize c1x1+.........+cnxn+0xn+1+........+0xm+n
Subject to a11x1+.........+a1nxn- xn+1 = b1
. . .
. . .
am1x1+.........+amnxn- xm+n = bm
x1 0, .........,xn 0, xn+1 0,.........,xn+m 0
(Only the dimensions of has been increased and ).
3. Describe the Two-Phase method of solving a linear programming problem with an
example.
The 2-Phase method is based on the following simple observation: Suppose that yo
u have a linear programming problem in canonical form and you wish to generate a
feasible solution (not necessarily optimal) such that a given variable, say x3,
is equal to zero. Then, all you have to do is solve the linear programming prob
lem obtained from the original problem by replacing the original objective funct
ion by x3 and setting opt=min.
If more than one variable is required to be equal to zero, then replace the orig
inal objective function by the sum of all the variables you want to set to zero.
Observe that because of the non-negativity constraint, the sum of any collection
of variables cannot be negative. Hence the smallest possible feasible value of
such a sum is zero. If the smallest feasible sum is strictly positive, then the
implication is that it is impossible to set all the designated variables to zero
.
To illustrate the modelling aspects of this approach let us re-examine the follo
wing little example:
x1 + 3x2 + 7x3 > = 25
-3x1 - 2x2 + 7x3 = - 5
2x1 + x2 + 4x3 < = 10
x2,x3 >= 0
Taking case of its violations of the standard form we obtain the following canon
ical form:
x1 - x2 + 3x3 + 7x4 - x5 +
x6 = 25
3x1 - 3x2 + 2x3 - 7x4 +
x7 = 5
2x1 - 2x2 + x3 + 4x4 +
x8 = 10
xj>= 0 , j=1,...,8
There are two artificial variables, namey x6 and x7. Thus, Phase 1 of the 2-Phas
e method involves the following linear programming problem:
w* := min w := x6 + x7
x1 - x2 + 3x3 + 7x4 - x5 +
x6 = 25
3x1 - 3x2 + 2x3 - 7x4 +
x7 = 5
2x1 - 2x2 + x3 + 4x4 +
x8 = 10
xj>= 0 , j=1,...,8
If we now incorporate the objective function in the constraint in the usual mann
er and place the new variable, namely w last, we obtain the following system:
x1 - x2 + 3x3 + 7x4 - x5 +
x6 = 25
3x1 - 3x2 + 2x3 - 7x4 +
x7 = 5
2x1 - 2x2 + x3 + 4x4 +
x8 = 10
-
x6 - x7 + w = 0
xj>= 0 , j=1,...,8, w>= 0
Note that this system is not in a canonical form because the columns of the arti
ficial variables are not elementary columns. To obtain a canonical form we have
to add the rows of the artificial variables to the last row. This yields:
x1 - x2 + 3x3 + 7x4 - x5 +
x6 = 25
3x1 - 3x2 + 2x3 - 7x4 +
x7 = 5
2x1 - 2x2 + x3 + 4x4 +
x8 = 10
4x1 - 4x2 + 5x3 + - x5 +
w = 30
xj>= 0 , j=1,...,8, w>= 0
This is then the system that will be used to initialise the simplex algorithm fo
r Phase 1 of the 2-Phase method. Of course, the column of w will not appear in t
he tableau.
We can ditinguish between two cases as far as the end of Phase 1 is concerned, n
amely:
Case 1: w* > 0 : The optimal value of w is greater than zero.
Case 2: w* = 0 : The optimal value of w is equal to zero.
In Case 1 we conclude that the LP problem under consideration does not have a fe
asible solution whereas Case 2 implies that the constraints are feasible, hence
the problem under consideration possesses a feasible solution.
The following final simplex tableau is an example of Case 1:
x1 x2 x3 x4 x5 x6 x7 w RHS
1 -1 2 7 -1 2 25
-3 1 -7 1 1 5
-2 -1 -4 -1 1 5
One artificial variable (x6) is in the basis and is not equal to zero. The probl
em does not have a feasible solution.
It should be noted that although Case 2 implies that all the artificial variable
s are equal to zero, this does not mean that they are all out of the basis.
So it is necessary to consider Case 1 in more detail, namey:
Case 2.1: All the artificial variables are non-basic.
Case 2.2: some of the artificial variables are in the basis
In Case 2.1 we proceed to Phase 2 of the 2-Phase method replacing the objecitve
function of Phase 1 with the original objecitve function. This will typically vi
olate the canonical form of the problem and thus pivot operations may have to be
used to restore the canonical form.
The following is a typical example of Case 2.1
x1 x2 x3 x4 x5 x6 x7 w RHS
1 -1 2 7 2 25
-3 1 -7 1 1 1 5
-2 -1 -3 -1 1 0
The two artificial variables are not in the basis, hence w is equal to zero.
Case 2.2 represents a degenerate basis, namely a situation where one or more of
the basic variables are equal to zero. Here is an example:
x1 x2 x3 x4 x5 x6 x7 w RHS
1 -1 2 7 -1 2 25
-3 1 -7 1 1 0
-2 -1 -4 -1 1 0
To take a degenerate artificial variable out of the basis we pivot on any non-ar
tificial variable whose coefficient in the row of the artificial is not equal to
 zero and enter it into the basis.
Fore example, in the table above, we can replace x6 by either x2 or x3 or x4.
If the coefficients of all the non-artificial variables in that row are zeros, t
hen the conclusion is that the constraint is redundant and thus can be ignored.

4. What do you understand by the transportation problem? What is the basic assum
ption behind the transportation problem? Describe the MODI method of solving tra
nsportation problem.
Transportation Problem
A typical transportation problem is shown in Fig. 9. It deals with sources where
a supply of some commodity is available and destinations where the commodity is
demanded. The classic statement of the transportation problem uses a matrix wit
h the rows representing sources and columns representing destinations. The algor
ithms for solving the problem are based on this matrix representation. The costs
of shipping from sources to destinations are indicated by the entries in the ma
trix. If shipment is impossible between a given source and destination, a large
cost of M is entered. This discourages the solution from using such cells. Suppl
ies and demands are shown along the margins of the matrix. As in the example, th
e classic transportation problem has total supply equal to total demand.
Matrix model of a transportation problem.
The network model of the transportation problem is shown in Fig. 10. Sources are
identified as the nodes on the left and destinations on the right. Allowable sh
ipping links are shown as arcs, while disallowed links are not included.
Network flow model of the transportation problem.
Only arc costs are shown in the network model, as these are the only relevant pa
rameters. All other parameters are set to the default values. The network has a
special form important in graph theory; it is called a bipartite network since t
he nodes can be divided into two parts with all arcs going from one part to the
other.
On each supply node the positive external flow indicates supply flow entering th
e network. On each destination node a demand is a negative fixed external flow i
ndicating that this amount must leave the network. The optimum solution for the
example is shown in Fig. 11.
. Optimum solution, z = 46.
Variations of the classical transportation problem are easily handled by modific
ations of the network model. If links have finite capacity, the arc upper bounds
can be made finite. If supplies represent raw materials that are transformed in
to products at the sources and the demands are in units of product, the gain fac
tors can be used to represent transformation efficiency at each source. If some
minimal flow is required in certain links, arc lower bounds can be set to nonzer
o values.

5. Describe the North-West Corner rule for finding the initial basic feasible so
lution in the transportation problem.
North West Corner Rule
The North West corner rule is a method for computing a basic feasible sol
ution of a transportation problem, where the basic variables are selected fro
m the North – West corner ( i.e., top left corner ). The standard North West Co
rner Rule instructions are paraphrased below:
Steps
• Select the north west (upper left-hand) corner cell of the transportati
on table and allocate as many units as possible equal to the minimum between ava
ilable supply and demand, i.e., min(s1, d1).
• Adjust the supply and demand numbers in the respective rows and columns.
• If the demand for the first cell is satisfied, then move horizontally to
the next cell in the second column.
• If the supply for the first row is exhausted, then move down to the fir
st cell in the second row.
• If for any cell, supply equals demand, then the next allocation can be m
ade in cell either in the next row or column.
• Continue the process until all supply and demand values are exhausted.
A procedure to find a basic and feasible solution for a transportation problem i
s North-West Corner rule.
Start in the upper left-hand (North West) corner of the tableau and assi
gn the most you can to the variable (considering supply and demand).
You then move down or to the right depending on supply and demand until
you reach the south east corner.
Constraints are equal to number of origins + number of destinations – th
e redundant constraint.
3 Factories + 4 Distributors –1 = 6 constraints.

6. Describe the Branch and Bound Technique to solve an I.P.P. problem.

The branch-and-bound technique can be outlined in simple terms. An enumeration t
ree of continuous linear programs is formed, in which each problem has the same
constraints and objective as except for some additional bounds on certain compon
ents of . At the root of this tree is the problem with the requirement removed.
The solution to this root problem will not, in general, have all integer compo
nents. We now choose some noninteger solution component and define to be the int
eger part .This gives rise to two subproblems. The left-child problem has the
additional constraint whereas in the right-child problem we impose . This branch
ing process can be carried out recursively; each of the two new problems will gi
ve rise to two more problems when we branch on one of the noninteger components
of their solution. It follows from this construction that the enumeration tree i
s binary.
Branch and bound (BB) is a general algorithm for finding optimal solutions of va
rious optimization problems, especially in discrete and combinatorial optimizati
on. It consists of a systematic enumeration of all candidate solutions, where la
rge subsets of fruitless candidates are discarded en masse, by using upper and l
ower estimated bounds of the quantity being optimized.
For definiteness, we assume that the goal is to find the minimum value of a func
tion f(x) (e.g., the cost of manufacturing a certain product), where x ranges ov
er some set S of admissible or candidate solutions (the search space or feasible
region). Note that one can find the maximum value of f(x) by finding the minimu
m of g(x) = − f(x).
A branch
and
bound procedure requires two tools. The first one is a splitting pr
ocedure that, given a set S of candidates, returns two or more smaller sets whos
e union covers S. Note that the minimum of f(x) over S is , where each vi is the
minimum of f(x) within Si. This step is called branching, since its recursive a
pplication defines a tree structure (the search tree) whose nodes are the subset
s of S. Another tool is a procedure that computes upper and lower bounds for the
minimum value of f(x) within a given subset S. This step is called bounding.
The key idea of the BB algorithm is: if the lower bound for some tree node (set
of candidates) A is greater than the upper bound for some other node B, then A m
ay be safely discarded from the search. This step is called pruning, and is usua
lly implemented by maintaining a global variable m (shared among all nodes of th
e tree) that records the minimum upper bound seen among all subregions examined
so far. Any node whose lower bound is greater than m can be discarded.
The recursion stops when the current candidate set S is reduced to a single elem
ent; or also when the upper bound for set S matches the lower bound. Either way,
any element of S will be a minimum of the function within S.