OR PPT 2

Chapter Three
Duality and Sensitivity Analysis
© 2006 Prentice Hall, Inc. B–1

DUALITY
 Duality is a property of the simplex method that adds further versatility to the
general LP model.
 The term dual indicates that there are two ways of looking at each problem.
Associated with every LPP, there is another intimately related LPP.
 The original LPP is called the primal problem while the corresponding
intimately related problem is called its dual problem.
 In fact either of the problems can be considered original as both originate

from the same data; consequently the other becomes its dual.
 Every LPP has another LPP associated with it, which is called its dual.
 The first way of starting a linear problem is called the primal of the problem.
 The second way of starting the same problem is called the dual.

 The optimal solutions for the primal and the dual are equivalent, but they are
derived through alternative procedures.
 The dual contains economic information useful to management, and it may also
be easier to solve, in terms of less computation, than the primal problem.
 Corresponding to every LPP, there is another LPP.
 The given problem is called the primal.
 The related problem to the given problem is known as the dual.
 The dual of a normal max problem is a normal min problem.
 Normal max problem is a problem in which all the variables are required to be
non-negative and all the constraints are ≤ constraints.

 Normal min problem is a problem in which all the variables are required to be
non-negative and all the constraints are ≥ constraints.
 The dual of a dual is the primal (original LP).
 If the primal has optimal solution, the dual will have optimal solution.
 If the primal has no optimal solution, the dual will not have optimal solution.
 Whether we follow the dual or primal system, the optimal solution will
remain equal.

The rules for constructing the dual from the primal or primal from the dual
are:
1. The number of variables in the dual problem is equal to the number of

constraints in the original problem. The number of constraints in the dual
problem is equal to the number of variables in the original problem.
2. Coefficients of the objective function in the dual problem come from the right
hand side of the original problem.
3. If the original problem is a maximization problem, the dual is a minimization

problem. If the original problem is a minimization problem, the dual is a
maximization problem.
4. The coefficients for the first constraint function for the dual problem are the
coefficients of the first variable in the constraints for the original problem, and
similarly for the other constraints.
5. The right hand sides of the dual constraints come from the objective
function coefficients in the original problem.
6. The dual of the dual problem is the original LPP itself.

Here we should note some of the symmetrical features as follows:

Here we should note some of the symmetrical features as follow:
The primal is maximization The dual is minimization and vice versa

In primal objective function written as Z In dual objective function written Z*
In primal the variables are X1 and X2 In dual we have new set of variables Y1, Y2
and Y3
Primal has two variables X1 and X2 The dual, therefore, has two constraints
The primal has three constraints The dual, therefore, has three variables
In primal’s objective function, 6 and 4 are In dual 6 and 4 become constants of
the coefficients constraints on the right hand side.
In primal the signs of constraints are less In dual the signs of constraints are just the
than or equal to ( ≤ ). reverse, greater than or equal to ( ≥ ).
The non-negativity constraints are as The non-negativity constraints are as
many as the variables in the primal, i.e. 2 many as the variables in the dual, i.e. 3
 The economic interpretation of primal and dual problems:
 The primal problem: with a given unit of value of each output (Cj) and a given
upper limit for the availability of each input (bi), the value of the total output.
 The dual problem: with a given availability of each input (bi) and a given lower
limit of unit value for each output (Cj), what unit values should be assigned to
each input in order to minimize the value of total input.
 The optimal values of the dual variables are also called shadow prices,
marginal value or opportunity cost.
 Duality Advantage:
1. The dual form provides an alternative form.

2. The dual reduces the computational difficulties associated with some formulation.
3. The dual provides an important economic interpretation concerning the value of
scarce resources used.

Example 1:
Max. Z = 5x1+ 6x2

Subject to:
2x1+3x2 < 3000 (Labor constraint)
5x1+7x2 < 1000 (Machine constraint)
x1 + x2 < 5000 (Market constraint)
x1, x2 > 0
Max. Z = 5x1+6x2
2x1+3x2 < 3000
5x1 + 7x2 < 1000
x1 + x2 < 5000

Example 1: Max. Z = 5x1+6x2
St 2x1+ 3x2 < 3000
5x1 + 7x2 < 1000
x1 + x2 < 5000
Represent primal in the conventional table as follows:
 Referring the above table, dual for this LP can be stated as:
Min Z* = 3000 u1 +1000 u2 +5000 u3
St. 2u1+5u2 + u3 > 5
u1+7u2 + u3 > 6
u1, u2 , u3 > 0
Note:
 For maximizing, all constraints must be brought to “<” form
 For minimizing, all constraints must be brought to “>” form
 If they are not, use multiplication factor -1
 “=” is an intersection of “>” and “< “
© 2006 Prentice Hall, Inc. B – 10
Example 2: Relationship of Primal- Dual problems
Max. Z = 60x1+50x2 Dual form is Min Z* = 80 u1 + 60u2 +15u3+36u4
St. St.
2x1+4x2 < 80 2u1+3u2 + u3 > 60
3x1 + 2x2 < 60 4u1+2u2 + u4 > 50
x1 < 15 u1, u2 , u3, u4 > 0
2x2 < 36
x1, x2 > 0
Find dual form

Solution
Primal is represented in the table as follows:

Example 3: Relationship of Primal- Dual problems
Primal Problem Max. Z = 2X1+ X2

St. - X1+ 2X2 < 2
X1+ X2 < 4
X1 < 3
X1, X 2 > 0
Standard Form:
Max. Z = 2X1+ X2 + 0S1 + 0S2 + 0S3
St. - X1+ 2X2 + S1 = 2
X1+ X2 + S2 = 4
X1 + S3 = 3
X1, X2 , S1, S2, S3 > 0

Max. Z = 2X1+ X2 + 0S1 + 0S2 + 0S3 St. - X1+ 2X2 + S1 = 2 Labor
X1+ X2 + S2 = 4 Material
Cj 2 1 0 0 0 X1 + S3 = 3 Capital
It. CB Bv X1 X2 S1 S2 S3 RHS (b) Ratio X1, X2 , S1, S2, S3 > 0
0 S1 -1 2 1 0 0 2 NA  Ratio = b/PR
0 0 S2 1 1 0 1 0 4 4
0 S3 1 0 0 0 1 3 3  EV= high +ve of Δj -- Max
Zi 0 0 0 0 0 0  LV= b/PR = smallest positive
Cj-Zi 2 1 0 0 0
 New EPR= LR /PN
0 S1 0 2 1 0 2 5 5/2
1 0 S2 0 1 0 1 -1 1 1  New R = OR - CPC(NER)
2 X1 1 0 0 0 1 3 NA
Zi 2 0 0 0 2 6  Zj = ΣCB.Xj
Cj-Zi 0 1 0 0 -2
0 S1 0 0 1 -2 3 3  Max Z = 2X1+X2 =2(3) + 1= 7
2 1 X2 0 1 0 1 -1 1  - X1+ 2X2 = -1 + 2(1) = 1; un
2 X1 1 0 0 0 1 3 used resource 1 unit and used is
Zi 2 1 0 1 1 7 1 of 2.
Cj-Zi 0 0 0 -1 -1  X1+ X2 = 4 (3+1) no un used
resource of 4; 4 units are used.
For Maximization -- Cj - Zi ≤ O -- Optimal solution.  X =3; no un used resource of 3.
1
Primal Problem Max. Z = 2X1+ X2
St. - X1+ 2X2 < 2
X1+ X2 < 4
X1 < 3
X1, X 2 > 0
Dual Problem Min Z* = 2y1 + 4y2 + 3y3

St - Y1 + y2 + y3 ≥ 2
2y1 + y2 ≥ 1
y1, y2, y3 ≥ 0
 Min = - Max
Standard form:
Max Z* = -2y1 - 4y2 - 3y3 - 0S1- 0S2- MA1-MA2

St - Y1 + y2 + y3 - S1 +A1 = 2
2y1 + y2 – S2 + A2 = 1
y1, y2, y3, S1, S2, A1, A2 ≥ 0

Max Z* = -2y1 - 4y2 - 3y3 + MA1-MA2  S1 and S2 have -ve value. Thus, they do not
St - Y1 + y2 + y3 - S1 +A1 = 2 considered as basic variable in the initial T.
2y1 + y2 - S2 + A2 = 1  Art. variables : once leave the process they
y1, y2, y3, S1, S2, A1, A2 ≥ 0 not considered again.
Cj -2 -4 -3 0 0 -M -M
It CB Bv y1 y2 y3 S1 S2 A1 A2 RHS Ratio
-M A1 -1 1 1 -1 0 1 0 2 2
0 -M A2 2 1 0 0 -1 0 1 1 1
Zi -M -2M -M M M -M -M -3M
Cj -Zi M-2 2M-4 M-3 -M -M 0 0
-M A1 -3 0 1 -1 1 1 1 1
1 -4 Y2 2 1 0 0 -1 0 1 NA
Zi 3M-8 -4 -M M -M+4 -M -M-4
Cj -Zi -3M+6 0 M-3 -M M-4 0
-3 Y3 -3 0 1 -1 1 1
2 -4 Y2 2 1 0 0 -1 1
Zi 1 -4 -3 3 1 -7
Cj -Zi -3 0 0 -3 -1
 For Maximization when Cj - Zi ≤ O (non-positive) = Optimal solution.

 No positive means addition of further variables can not improve net profit.
 Min = - Max. Thus, Min = -(-7) = 7
© 2006 Prentice Hall, Inc.  Max Z* = -2y1 - 4y2 - 3y3 = -2(0) – 4(1) – 3(1) = -7 = -(-7) =7 B – 15
Primal - optimal iteration (iteration 2)
Cj 2 1 0 0 0 Optimal solution
It. CB Bv X1 X2 S1 S2 S3 RHS Ratio
0 S1 0 0 1 -2 3 3
2 1 X2 0 1 0 1 -1 1
2 X1 1 0 0 0 1 3
Zi 2 1 0 1 1 7
Cj -Zi 0 0 0 -1 -1
Dual - optimal iteration (iteration 2)

Cj -2 -4 -3 0 0 -M -M Optimal solution
It CB Bv y1 y2 y3 S1 S2 A1 A2 RHS Ratio
-3 Y3 -3 0 1 -1 1 1
2 -4 Y2 2 1 0 0 -1 1
Zi 1 -4 -3 3 1 -7
Cj -Zi -3 0 0 -3 -1
 Primal S1 value = Dual Y1 value --- -(S1Cj-Zi) = -(0) = o then y1 = 0
 Primal S2 value = Dual Y2 value --- -(S2 Cj-Zi) = -(-1) = 1 then y2 = 1
 Primal S1 value = Dual Y1 value --- -(S3 Cj-Zi) = -(-1) = o then y3 = 1
 Dual S1 value = Primal X1 value --- -(S1 Cj-Zi) = -(-3) = 3 then X1 = 3
 Dual S 2 value = Primal X2 value --- -(S2 Cj-Zi) = -(-1) = 1 then X2 = 1
Sensitivity Analysis
 One of the assumptions of a LPP is the assumption of certainty.
 This assumption implies that the coefficients of a LPP are completely known
(determined) and do not change during the period being studied.
 That means, Cj (the per unit (profit or cost) contribution of each decision
variable), bi (the availability of resources), and aij is the technical coefficients
(per unit resource consumption or production of each decision variables) are
constants and known with certainty.
 However, in realty these coefficients are subject to change with time or error.
 If such changes will be occurred, there should be a means to check for
how long the present optimal solution continues as optimal.
o The method of evaluating the degree to which the present optimal solution is
continued as optimal is called sensitivity analysis.
17
 Sensitivity Analysis is concerned with the study of ‘Sensitivity ‘of the optimal
solution of LPP with decision variables (changes) in parameters.
 The degree of sensitivity of the solution due to those variations can range from
No change at all to a substantial change in the optimal solution of the given LPP.
 The process of studying the sensitivity of the optimal solution of a LPP is

called post-optimal analysis.
The two sensitivity analysis approaches are:
I. Trial and error approach and
II. Analytical approach

 Five types of discrete changes in the original LP model may be investigated
during the sensitivity analysis:
 Changes of the coefficients of the objective function (cj)

 Changes of the RHS Quantity( bj)
 Changes of the input-output coefficient
 Add/delete constraints
 The addition of a new variable to the problem
19
 Changes of the coefficients of the objective function (cj)
Decision variables can be:
 Basic (in the solution)
 Non-basic (out-of the solution)
 Changes in Cj of basic variables
 One of the important parameters of LPP is the coefficient of objective function.
 The test of sensitivity of the objective function involves finding the range of
values with in which each Cj can lie without changing the current optimal
solution.
20
 Change in coefficient Cj of non-basic variable Xj
 Among the variables included in the LPP some are non - basic variables.
 If Cj is the coefficient of non-basic variable Xj:

 It doesn’t affect any of Cj values of basic variables.
 It doesn’t affect any of Zj values.
 It affects ∆j values.
 If optimality is to be maintained for maximization LPP, all coefficients in the

index row (Δj = Cj - Zj ) should be non-negative (positive or zero).
 Note: Instead of resolving the entire problem as a new problem with new
parameters, we may take the original optimal solution table as an initial solution
table for the purpose of knowing ranges both lower and upper within which a
parameter may assume value.

 Range for the coefficients of basic decision variables
 The range of optimality is the range over which a basic decision variable
coefficient in the objective function can change without changing the optimal
solution mix.
 However, this change will change only the optimal value of the objective
function.

Example
Maximize Z = 60x1 + 50x2
S.t
4x1 + 10x2 < 100
2x1 + 1x2 < 22
3x1 + 3x2 < 39
x1, x2 ≥ 0
 The final simplex tableau for that problem is repeated here for convenience.
 Determine the range of optimality for coefficients of the basic decision variables.
23
Solution:
Analysis of basic decision variables:
The analysis will be conducted on products X1 and X2 which are in the basic
solution. Divide each Cj - Zj row entry for variables not in the solution (for instance,
by S2 and S3 values) by the associated variable aij from X1 or X2 row.
Analysis of X1
S2 S3
Cj – Zj row = A -10 -40/3
X1 row = B 1 -1/3
A/B -10 40
Steps:
1.Take the Cj - Zj row of the optimal solution of the non-basic variables.
2.Take the X1 row of the non-basic variables.
3. Cj - Zj row
X1 row
25
Analysis of X2

 The range for the non-basic variables
 If there is a variable Cj, not participating in the optimal basis, then, in order for
the variable to be included in the optimal solution, its coefficient in the objective
function will have to change from the existing Cj to a new level Cj (new).
 The range of insignificance is the range over which Cj rates for non-basic
variables can vary without causing a change in the optimal solution mix (variable)
is called the range of insignificance.

Chapter Four
Non-linear Programming

Introduction
 A key assumption of linear programming is that all its functions (objective

function and constraint functions) are linear.
 Although this assumption essentially holds for many practical problems, it

frequently does not hold.
 Therefore, it often is necessary to deal directly with nonlinear programming

problems, so we turn our attention to this important area.
 In one general form, the nonlinear programming problem is to find:
X = (X1, X2, . . . , Xn) so as to Maximize f(x),

Subject to
gi(x) ≤ bi, for i = 1, 2, . . . , m, and
x ≥ 0,
 Where f(x) and the gi(x) are given functions of the n decision variables.

 There are many different types of nonlinear programming problems, depending
on the characteristics of the f(x) and gi(x) functions.
 Different algorithms are used for the different types.
 For certain types where the functions have simple forms, problems can be
solved relatively efficiently.
 For some other types, solving even small problems is a real challenge.
 Because of the many types and the many algorithms, nonlinear programming is a
particularly large subject.
 We do not have the space to survey it completely.
 However, we do present a few sample applications and then introduce some of

the basic ideas for solving certain important types of nonlinear programming
problems.

 If a nonlinear programming problem has no constraints:
 The objective function being concave guarantees that a local maximum is a

global maximum.
 Similarly, the objective function being convex ensures that a local

minimum is a global minimum.
 If there are constraints:
 Then one more condition will provide this guarantee, namely, that the
feasible region is a convex set.
 For this reason, convex sets play a key role in nonlinear programming.

 A global maximum (absolute maximum): is the largest overall value of a set,
function etc. over its entire range.
 A local maximum point on a function is a point (x, y) on the graph of the

function whose y coordinate is higher than all other y coordinates on the graph at
points close to (x, y).
 To find local maximum or minimum, first, the first derivative of the function
needs to be found.
 Values of X which makes the first derivative equal to zero are critical points.
 If the second derivative at x = c is positive, then f(c) is a minimum.
 When the second derivative is a negative at x = c the f(c) is maximum.

Types of nonlinear programming problems
 Nonlinear programming problems can have different shapes and forms.
 Unlike the simplex method for linear programming, no single algorithm can
solve all these different types of problems.
 Instead, algorithms have been developed for various individual.
Maximize Z = 3X1 + 5X2

Subject to X1 ≤ 4
2X2 ≤ 14
8X1 - X12 + 14X2 - X22 ≤ 49
X1, X2 ≥ 0
Example: With 2X2 ≤ 14 and a nonlinear constraint (8X1 - X12 + 14X2 - X22 ≤ 49)
replacing the original second and third functional constraints.

X2
(0, 7) = Optimal solution
Z = 35 =3X1 + 5X2
3 (4, 3) = Local maximum
Z = 27 =3X1 + 5X2
4 X1

. Unconstrained Optimization
 Unconstrained optimization problems have no constraints, so the objective is
simply to Maximize f(x).
 Over all values of X = (X1, X2, . . . , Xn).
 The necessary condition that a particular solution X = X* be optimal when f(x)

is a differentiable function.
 𝑑𝑓 / 𝑑𝑥 = 0 at X = X*, for j = 1, 2, . . . , n.
 When f(x) is a concave function, this condition is also sufficient, then solving
for X* reduces to solving the system of n equations obtained by setting the n
partial derivatives equal to zero.
 Unfortunately, for nonlinear functions f(x), these equations often are going to be
nonlinear as well, in which case it is unlikely to be able to solved analytically for
their simultaneous solution.

 What are the procedures for finding X*: First for n = 1 and then for n > 1.
 These procedures also play an important role in solving many of the problem
types described next, where there are constraints.
 The reason is that many algorithms for constrained problems are designed so
that they can focus on an unconstrained version of the problem during a portion
of each iteration.
 When a variable Xj does have a non-negativity constraint Xj ≥ 0, the

preceding necessary and (perhaps) sufficient condition changes slightly to:
 𝑑𝑓/𝑑𝑥𝑗 ≤ 0 at X = X*, if Xj* = 0

= 0 at X = X* if Xj* > 0 for each such j.

 This condition is illustrated in Fig. below, where the optimal solution for a
problem with a single variable is at X = 0 even though the derivative is negative
(-2) rather than zero.
 Because this example has a concave function to be maximized subject to a non-

negativity constraint, having the derivative less than or equal to 0 at X = 0 is
both a necessary and sufficient condition for X = 0 to be optimal.
 A problem that has some non-negativity constraints but no functional

constraints is one type of special case (m = 0) of the problems.

Linearly Constrained Optimization
 Linearly constrained optimization problems are characterized by constraints that

completely fit linear programming, so that all the gi(x) constraint functions are
linear, but the objective function f(x) is nonlinear.
 The problem is considerably simplified by having just one nonlinear function to

take into account, along with a linear programming feasible region.
 A number of special algorithms based upon extending the simplex method to

consider the non-linear objective function have been developed.
 One important special case, is quadratic programming.

Quadratic Programming
 Quadratic programming problems have linear constraints, but the objective

function f(x) must be quadratic.
 Thus, the only difference between such a problem and a linear programming
problem is that some of the terms in the objective function involve the square
of a variable or the product of two variables.
Maximize f (x) = 24 - 2X - X2
Subject to
X≥0
 The example shows how an optimal solution can lie at a point where a derivative
is negative instead of zero, because that point lies at the boundary of a non-
negativity constraint.

f(x)
Global maximum because f(x) is concave and df / dx = -2
24
20
4 5 x

 Many algorithms have been developed for this case under the additional
assumption that f(x) is a concave function.
 Quadratic programming is very important, partially because such formulations

arise naturally in many applications.
 For example, the problem of portfolio selection with risky securities.
 However, another major reason for its importance is that a common approach to
solving general linearly constrained optimization problems is to solve a sequence
of quadratic programming approximations.

Convex Programming
 Convex programming covers a broad class of problems that actually

encompasses as special cases all the preceding types when f(x) is a concave
function to be maximized.
 Continuing to assume the general problem form (including maximization)

presented at the beginning of the chapter, the assumptions are that:
1. f(x) is a concave function.

2. Each gi(x) is a convex function
 When we maximize an ordinary (doubly differentiable) function of a single

variable f(x) without any constraints, this guarantee can be given when:
 𝑑𝑦2 / 𝑑𝑥2 ≤ 0 for all x.
 Such a function that is always “curving downward” (or not curving at all) is
called a concave function.

 Similarly, if ≤ is replaced by ≥, so that the function is always “curving upward”
(or not curving at all), it is called a convex function.
 Thus, a linear function is both concave and convex.
 Functions of multiple variables also can be characterized as concave or convex

if they always curve downward or curve upward.
 These intuitive definitions are restated in precise terms, along with further
elaboration on these concepts.

Separable Programming
 Separable programming is a special case of convex programming, where the

one additional assumption is that:
 3. All the f(x) and gi(x) functions are separable functions.
 A separable function is a function where each term involves just a single

variable, so that the function is separable into a sum of functions of
individual variables.
 For example, if f(x) is a separable function, it can be expressed as:

𝑛
𝑓(x) = ∑ 𝑓𝑗(X𝑗)
𝑗=1
 Where each fj(Xj) function separable programming problems satisfy the

assumption of additivity but violate the assumption of proportionality when any
of the fj(Xj) functions are nonlinear functions.

 It includes only the terms involving just Xj in the terminology of linear
programming.
 f(X1, X2) = 126X1 – 9X12 + 182X2 – 13X22
 It is a separable function because it can be expressed as:(X1, X2) = f1(X1) + f2(X2)
 Where:
 f1(X1) = 126X1 – 9X12 and

 f2(X2) = 182X2 – 13X22
o are each a function of a single variable X1 and X2, respectively.

Chapter Five
Risk Programming and Efficiency
Analysis
5.1. Risk in Agricultural Economics

5.2.Types of Risk Programming
5.3. Application of Game Theory
5.4. Efficiency Analysis Methods

5.1. Risk in Agricultural Economics

Introduction
 Risk is the possibility that events will occur and affect the achievement strategy
and business objectives.
 It is often considered in terms of severity.
 In some instances, risk could relate to the anticipation of an expected event that
does not occur.
 This means that an entity's strategy and business objectives may be affected
by potential events.
 The lack of predictability of an event occurring (or not) and its related impact
creates uncertainty for an organization.
 It is then understood that any uncertainty exists for entities that set out to achieve
future strategies and business objectives.

 Agriculture as a business is often carried out in the open air, and always entails
the management of inherently variable living plants and animals, it is especially
exposed to risk.
 Different forms of risk such as production risk, price or market risk, institutional
risk, personal or human risk, business risk and financial risk are risks commonly
take place in agricultural activities.
 Production risk comes from the unpredictable nature of the weather and
uncertainty about the performance of crops or livestock, eg through the incidence
of pests and diseases or from many other unpredictable factors.
 Prices of farms inputs and outputs are seldom known for certain at the time that a
farmer must make decisions about how much of which inputs to use or what and
how much of varies products to produce, but it can be changed due to unexpected
cases.
 Risk in agriculture matters due to:

 Most people dislike risk and
 Downside risk.
 Most people dislike risk:
 Most people are risk averse when faced with significantly risky incomes or
wealth outcomes.
 A person who is risk averse will be willing to forgo some expected return for a
reduction in risk, the rate of acceptable trade-off depending on how risk averse
that individual is.
 The existence of risk aversion means that analysis of risky choices interims of
their average or expected consequences will not always lead to the identification
of the option that will be most preferred.

 Downside risk
 The great bulk of analytical investigation of risk in investment appraisal and

project evaluation has been concerned with ‘pure risk’, that is, variation about
some measure of ‘ representative’ performance that is the focus of attention, such
as the mean.
 Down side risk is more relevant in a variety of situations including many analysis
of risk in agriculture.
 Downside risk is the shorthand description for those situations in which any
significant deviations from the ‘norm’ lead to worse outcomes.
 Downside risk refers to the probability that an asset or security will fall in price.

Types of Risk Programming
Linear risk programming

 Linear programming is the widely used method for farm planning.
 In the context of the whole farm planning under risk, LP may be used to
maximize the expected profit subject to the farm resource constraints and other
restrictions.
 The notion is as follows:
Maximize E = CX - f
St
AX ≤ b ---- Canonical form of Maximization
and
X≥0
E(X) = E = Σ P(Xi)
Where: E = is expected profit C = a 1xn vector of activity

X = is nx 1vector of activity levels F = is a fixed overhead costs
Non-linear risk programming
 No specific mathematical procedure followed in order to find optimal solution for

non linear risk programing as in the case of linear programming.
 General form: Maximize f(x): Subject to:

gi (x) ≤ bi
and
X≥0
 Where f(x) and g(x) are given functions of n decision variables.
 There is no algorithm to solve this general form of non-linear models.
 But different non-linear problems have their own forms that can solve specific
problems fitting with their format.

 Types of Non-linear risk programing problems are:
 Linear constrained optimization

 Quadratic programing
 Convex programing
 Concave Programming
 Separable programing
 Fractional programing

5.2. Game Theory

History of game theory
 The individual closely associated with the creation of the theory of games is John
von Neumann, one of the greatest mathematicians of this century. Although
others proceeded him in formulating a theory of games - notably Emile Borel - it
was von Neumann who published in 1928 the paper that laid the foundation for
the theory of two-person zero-sum games.
 The theory of Games was born in 1944 with the publication of Theory of Games
and Economic Behaviour by Hungarian-born American mathematician John von
Neumann and his Princeton University colleague Oskar Morgenstern, a German-
born American economist. In their book, . They observed that economics is
much like a game, wherein players anticipate each other’s moves, and
therefore requires a new kind of mathematics, which they called game theory.
Their choice of title was a little unfortunate, since it quickly got shortened to
“Game Theory,”
 Nobel Laureate and a Father of Game Theory, Lloyd S. Shapley(92), who shared
the 2012 Nobel Memorial Prize in Economic Science for work on game theory
that has been used to study subjects as diverse as matching couples and
allocating costs.
Application of Game Theory
Overview of Game Theory
 Game theory is an approach to modeling behavior in situations where the

outcome of your decisions depends on the decisions of others.
 Game theory is the study of strategic, interactive decision making among

rational individuals or organizations.
 Game theory is a branch of applied mathematics that provides tools for analyzing
situations in which parties, called players, make decisions that are interdependent.
 This interdependence causes each player to consider the other player’s possible
decisions (or strategies) in formulating strategy.
 In addition, a player need not be an individual; it may be a nation, a corporation,

or a team comprising many people with shared interests.

 A solution to a game describes the optimal decisions of the players, who may
have similar, opposed, or mixed interests, and the outcomes that may result from
these decisions.
 Game theory is applied for determining different strategies in the business world.
 It offers valuable tools for solving strategy problems.

Classic examples of game theory
1. The Prisoner’s Dilemma:
 Where two suspects are in police custody (keeping) as accomplices for the same
crime, but there is not enough evidence for a crime conviction (belief).
 They are held and interrogated (interviewed) separately.
 If one prisoner testifies (witnessed) against the other while the other stays silent,
the testifying prisoner goes free and the silent prisoner is convicted (imprisoned)
and serves ten years.
 If both prisoners stay silent, both are imprisoned on a minor charge and serve six
years.
 If both prisoners testify against each other, each serves five years.
 How should the prisoners act?
 The answer is that both prisoners should testify against the other, an outcome that
is known as a Nash equilibrium.
 The players in each game depend on each other’s rationality to make the
optimal choices in every situation, thereby maximizing their own utility.
2. Conflict of sales
 Let us assume that the client had a drug with dominant market share in a
therapeutic area that is competitive, yet has few players.
 One of their competitors was behaving aggressively from a pricing perspective

(e.g., continually matching and beating the client’s price, even though the
competitor gained little or no market share from such actions).
 The competitor was in the process of running a clinical trial to improve their
product’s label.
 The client wished to know how the competitor might behave depending on the
client’s next price move and the outcome of the clinical trial.
 The conclusion of the exercise was that engaging in a price war would not benefit
any of the client or competitors.
 Therefore, keeping the price high and trading “market share for peace” would be
the most profitable strategy.
3. Advertising War:
 Coke Vs Pepsi
 Without any advertising, each company earns $5b/year from Cola consumers.
 Each company can choose to spend $2b/year on advertising.
 Advertising does not increase total sales for Cola, but if one company advertises
while the other does not, it captures $3b from the competitor
Pepsi
No Ad Ad
No Ad 5birr 5 birr 2birr 6birr
Coke
Coke Ad 6 birr 2 birr 3birr 3birr
 What will the Cola companies do? Is there a better feasible outcome.

4. In your everyday life:
 Everything is a game, poker, chess, soccer, driving, dating, stock market

advertising, setting prices, entering new markets, building a reputation
bargaining, partnerships, job market search and screening designing contracts,
auctions, insurance, environmental regulations international relations, trade
agreements, electoral campaigns, Most modern economic research includes game
theoretical elements.
5. Game theory has been used to analyze parlour games, but its applications are
much broader.

Classification of game theory
 It broadly classified into three main sub-categories of study.
1. Classical game theory
 It focuses on optimal play in situations where one or more people must make a
decision and the impact of that decision and the decisions of those involved is
known.
 Decisions may be made by use of a randomizing device like piping a coin.
 It focuses on questions like:
 What is my best decision in a given economic scenario?

 Where a reward function provides a way for me to understand how my
decision will impact my result.
Examples: Poker, Strategic military decision making, Negotiations.

2. Combinatorial game theory
 It focuses on optimal play in two-player games in which each player takes turns
changing in pre-defined ways.
 In other word, combinatorial game theory does not consider games with chance
(no randomness).
 Generally two player strategic games played on boards.
 Moves change the structure of a game board.
Examples: Chess, Checkers, Go.

3. Dynamic game theory:
 It focuses on the analysis of games in which players must make decisions over
time and in which those decisions will affect the outcome at the next moment
in time.
 It often relies on differential equations to model the behavior of players over time.
 It can help optimize the behavior of unmanned vehicles or it can help you capture
your baby sister who has escaped from her playpen.
 In general games with time, Games with motion or a dynamic component.
Examples: Optimal play in a dog fight, Chasing your brother across a room.

Key elements in game theory
 Player: who is interacting
 Strategies: what are the options of each player? In what order do players act?
 Payoffs: How do strategies translate into outcomes? What are players’ preferences
over possible outcomes?
 Information/Beliefs: What do players know/believe about the situation and about

one another?
What actions do they observe before making decisions?
 Rationality: How do players think?

Mathematical concept in Game Theory
 The following the prerequisite required for game theory
Sum of gains and loss:
 If in a game sum of the gains to one player is exactly equal to the sum of losses to
another player, so that sum of the gains and losses equal zero, the corresponding
game is said to be zero sum game.
Types of games
 Games can be classified according to certain significant features, the most obvious
of which is the number of players.
 Thus, a game can be designated as being a one-person, two-person, or n-person

(with n greater than two) game, games in each category having their own
distinctive features.

One-person games
 One-person games are also known as games against nature.
 With no opponents, the player only needs to list available options and then
choose the optimal outcome.
 When chance is involved the game might seem to be more complicated, but in
principle the decision is still relatively simple.
Example: Aperson deciding whether to carry an umbrella. While this person may
make the wrong decision, there does not exist a conscious opponent.
 That is, nature is presumed to be completely indifferent to the player’s decision,

and the person can base his decision on simple probabilities.
 One-person games hold little interest for game theorists.

Two person zero sum game (with two players)
 The game in which there are exactly two player and the interest of the players
completely opposed are referred as two-person zero sum games.
 They are called zero sum games because one player wins whatever the other
player loses.
 In short it is denoted by TPZS game.
Example: All parlour game and sports, like Tic-tac-toe, chess, cribbage,
backgammon, and tennis etc., are TPZS games
Two person zero sum game (with more than two players)
 TPZS games with more than two people involved are:
1. Team sports with only two sides, but with more than one player in each side.
2. Many people involved in surrogates for military conflict, so it should come as no

surprise that many military problems can also be analyzed as TPZS games.
Games which are not TPZS
1. Those parlour games in which the players cannot be clearly separated into two
sides are not TPZS games.
2. Those poker and Monopoly games when played by more than two people are not
TPZS games.
3. Most real economic “games” are not TPZS, because there are too many players,
and also the interests of the players are not completely opposed.
Positive-sum game
 In game theory, a term positive sum refers to situations in which the total of
gains and losses is greater than zero.
 A positive sum occurs when resources are somehow increased and an approach is
formulated such that the desires and needs of all concerned are satisfied.

Perfect games
 Games of perfect information in which each player knows everything about the
game at all times; it is called perfect games.
Example: Chess in which each player knows everything about the game at all times.
In chess exactly one of three outcomes must occur if the players make
optimal choices:
1. White wins (has a strategy that wins against any strategy of black),
2. Black wins
3. White and black draw.
Imperfect games
 Games of imperfect information in which each player do not knows everything

about the game at all times; it is also called imperfect games.
Example: Poker in which players do not know all of their opponents’ cards.

Finite games
 Games in which each player has a finite number of options, the number of
players is finite, and the game cannot go on indefinitely.
Example: chess, checkers, poker, and most parlour games are finite.
Cooperative games
 In game theory, a cooperative game (or coalitional game) is a game with

competition between groups of players ("coalitions") due to the possibility of
external enforcement of cooperative behaviour (e.g. through contract law).
Non cooperative games
 Those are opposed to cooperative games in which there is either no possibility to

forge alliances or all agreements need to be self-enforcing (e.g. through credible
threats).
Pay off
 The outcome of the game resulting from a particular decision (or strategy) is
called pay off.
 It is assumed that pay off is also known to the player in advance.
 It is expressed in terms of:
 Numerical values such as money

 Percent of market share or utility.

Pay off matrix
 The pay offs in terms of gains or losses, when players select their particular
strategies, can be represented in the form of matrix is called pay off matrix.
 Let A1, A2,....Am are possible strategies for player A.

 Let B1,B2,....Bn are possible strategies for player B.
 The total number of possible outcomes are m x n and it is assumed that each
player knows not only his own list of possible course of action but also his
opponent.
 For our convenience, we assume that player A always a gainer whereas player B
a looser.
 Let aij = pay off which player A gain from player B, if player A choose strategy
i and player B chooses strategy j.

Pay off matrix of player A is represented in the form of table
Player B’s strategies
Player A’s strategies B1 B2 --- Bn
A1 a11 A12 --- A1n
A2 A21 A22 --- A2n
-- --- --- --- ---
Am Am1 Am2 -- amn
Remarks
 For the zero sum games, the gain of one player is equal to the loss of other and
vice versa.
 i.e one player pay off table would contain the same amounts in pay off table of
other player with the sign changed.
 Therefore it is enough to construct pay of table for one player.

Strategy
 The strategy for a player is the list of all possible actions (moves or course of
action) that he will take for every pay-off (outcome) that might arise.
 It is assumed that all course of possible actions are known in advance to the
player.
Types of Strategy
 Usually player in game theory uses two types of strategy namely pure and mixed
strategy .
Pure strategy:
 Particular course of action that are selected by a player is called pure strategy
(course of action).
 i.e: each player knows in advance of all strategies out of which he always selects
only one particular strategy regardless of the other players strategy, and objective
of the player is to maximize gain or minimize loss
2. Mixed strategy:
 Course of action that are to be selected on a particular occasion with some

fixed probability are called mixed strategies.
 i.e: There is a probabilistic situation and objective of the players is to maximize

expected gain or minimize expected losses by making choice among pure
strategy with fixed probabilities.
 In mixed strategy, If there are 'n' number of pure strategies of the player, there
exist a set S ={p1,p2,.....pn}.
 Where pj is the probability with which the pure strategy j would be selected
and whose sum is unity.
 i.e: p1+p2+.....+ pn =1 and pj ≥ 0 for all j =1, 2,...n.

Remark:
1. If a player randomly chooses a pure strategy, we say that the player is using a
"mixed strategy."
 In a pure strategy a player chooses an action for sure, whereas in a mixed

strategy, he chooses a probability distribution over the set of actions available
to him.
2. If a particular Pj =1 and all others are zero, then the player is said to select pure
strategy j.
Optimal strategy
 The particular strategy (or complete plan) by which a player optimizes his gains
or losses without knowing the competitor's strategies is called optimal strategy.
Value of the game

 The expected outcome when players follow their optimal strategy is called the
value of the game, it is denoted by V.
Basic assumptions of game
1. Each player has available to him a finite number of possible strategies.
 The list may not be same for each player.
2. Player A attempts to maximize gains and player B minimize losses.
3. The decisions of both players are made individually prior to the play with no
communication between them.
4. The decisions are made simultaneously and also announced simultaneously

so that neither player has an advantage resulting from direct knowledge of the
other player 's decision.
5. Both the players know not only possible pay offs to themselves but also other.

Minmax-Maxmin principle
 The selection of an optimal strategy by each player without the knowledge of

the competitor's strategy is the basic problem of playing games.
 The objective of the study is to know how these players select their respective
strategy so that they may optimize their pay off.
 Such a decision making criterion is referred to as the minmax-maxmin principle.
Remarks
 Minmax-Maxmin principle give the best possible selection of strategy for both
players in pure strategy problem.

Saddle point
 If the min-max value = max-min value, then the game is said to have a saddle
(equilibrium) point.
Remarks
 The corresponding strategy at saddle point are called optimum strategy.
 The amount of pay off at an saddle point is known as the value of the game.
 A game may have more than one saddle point.
 There are also games without saddle point.
 Its name derives from its being the minimum of a row that is also the maximum
of a column in a payoff matrix which corresponds to the shape of a saddle.
Procedure to determine saddle point
Step 1. Select the minimum (lowest ) element in each row of the pay off matrix and
write them under 'row minima' heading.
 Then select the largest element among these elements and enclose it in a
rectangle.
Step 2. Select the maximum (largest) element in each column of the pay of matrix
and write them under 'column maxima' heading.
 Then select the lowest element among these elements and enclose it in a circle.
Step 3. Find out the elements which is same in the circle as well as rectangle and
mark the position of such elements in the matrix.
 This element represents the value of the game and is called the saddle point.

Games without saddle point
 Suppose if there is no pure strategy solution for a game, then there is no saddle
point exist.
 In these situations, to solve the game both the player must determine the
optimal mixtures of strategies to find saddle point.
 The optimum strategy mixture of each player may be determined by assigning

each strategy it probability of being chosen.
 The optimal strategy so determined is called mixed strategy.
Fair Game
 If the value of the game is zero (i.e. there is no loss or gain for any player), the
game is called fair game.
 The simplest type of game is one where the best strategies for both players are
pure strategies.
 This is the case if and only if, the pay-off matrix contains a saddle point.
Strictly determinable
 A game is said to be strictly determinable if the maxmin and minmax values of

the game are equal and both equal the value of the game.
Example: Games with sadle point
 Find the optimal plan for both the player
Player B
I II III IV
I -2 0 0 5
Palyer A
II 4 2 1 3
III -4 -3 0 -2
IV 5 3 -4 2

Solution:
 We use maxmin-minmax principle for solving the game.
Player B Row Minimum

I II III IV
I -2 0 0 5 -2
Player A II 4 2 1
3 1
III -4 -3 0 -2 -4
IV 5 3 -4 2 -4
Column Maximum 5 3 1 5
 Select minimum from the column maximum values.

ie. Minimax = 1, (marked as circle )
 Select maximum from the row minimum values

ie. Maximin = 1, ( marked as rectangle)

 Player A will choose strategy II, which yields the maximum payoff of 1.
 Player B will choose strategy III.
 The value of game is 1, which indicates that player A will gain 1 unit and player
B will sacrifice 1 unit.
 Since the maximin value = the minimax value =1, therefore the game has saddle
point and the game is not fair game (since value of the game is non zero).
 Also maxmin = minimax = value of game, therefore the game is strictly

determinable.
 It is a pure strategy game and the saddle point is (A-II, B-III)
 The optimal strategies for both players given by pure strategy:
 Player A must select strategy II and
 Player B must select strategy III.

Example 2: For the game with payoff matrix
Player B
Player A -1 2 -2
6 4 -6
 Determine the best strategies for players A and B and also the value of the game.
 Is this game (i) fair (ii) strictly determinable?
Solution:
Player B Row minimum
Player A -1 2 -2 -2
6 4 -6 -6
Column maximum 6 4 -2
 Select minimum from the column maximum values. ie. Minimax = -2, (marked as circle ).
 Select maximum from the row minimum values. ie. Maximin = -2, ( marked as rectangle).

 Player A will choose strategy I, which yields the maximum payoff of -2.
 Player B will choose strategy III.
 The value of game is -2, which indicates that player A will gain -2 unit and
player B will sacrifice -2 unit.
 Since the maximin value = the minimax value =-2, therefore the game has saddle
point and the game is not fair game (since value of the game is non zero)
 Also maxmin = minimax = value of game, therefore the game is strictly

determinable.
 It is a pure strategy game and the saddle point is (A-I, B-III)
 The optimal strategies for both players given by pure strategy, Player A must
select strategy II and player B must select strategy III.

Example 3: Find the range of values of p and q which will render the entry (2,2) a
sadle point of the game.
Row
Player B
minimum
Player A B1 B2 B3
A1 2 4 5 2
A2 10 7 Q 7
A3 4 P 6 4
Column 10 7 6
Maximum
 Here maxmin =7, minmax = 6, i.e the value of the game may be between 6 and 7.
 i.e maxmin is not equal to minmax, therefore there is no unique saddle point.
Games with no saddle point should solved using mixed strategy.
 If the saddle point exist at (2,2), only if q >7 and p ≥7.

5.3. Efficiency Analysis Methods

Introduction
 Efficiency can be expressed as a ratio by using output divided by input.
 Output, or work output, is the total amount of useful work completed with out
accounting for any waste and spoilage.
 Efficiency can also expressed as a percentage by multiplying the ratio by 100.
 Efficiency concerned with comparing the inputs with the outputs and determining
how best to maximize outputs for the resources allocated.
 The underlying aim of efficiency analysis is to understand how inputs are

translated in to value outputs.
 Efficiency is the value of output produced to the cost of inputs used in the
production process.

 Economists usually classifies efficiency in to three types namely:
 Technical efficiency
 Allocative efficiency and
 Economic efficiency.
Technical efficiency (TE): It is the attainment of the maximum potential yield from
specified quantities of inputs.
 TE is measured within two contexts: input-oriented and output-oriented.
 In an input-oriented framework, TE represents the potential input reduction that

a farm could implement without lowering output.
 In an output-oriented framework, TE gives information about the potential

output increase that a farm could implement without increasing its use of
inputs.
 In the case of constant returns to scale, both orientations give similar results.
 TE is defined as producing the maximum amount of output from a given amount
of input, or alternatively, producing a given output with the minimum input
quantities.
 It refers to the achievement of the maximum potential output from given

amounts of inputs, taking into account physical production relationships.
 A typical producer is technically efficient if an increase in any output requires a

reduction in at least one other output or an increase in at least one input, and if a
reduction in any input requires an increase in at least one other input or a
reduction in at least one output.
 Thus, a technically efficient producer could produce the same outputs with less of
at least one input, or could use the same inputs to produce more of at least one
output.

Allocative Efficiency (AE): Refers to the ability to combine inputs and outputs in
optimal proportions in the light of prevailing prices.
 AE shows whether the use of different proportions of production factors

guarantees the attainment of maximum production with a particular market price.
 Unlike TE concept that only considers the process of production, AE concept

pertains to the idea that society is concerned with not only how an output is
produced, but also with what outputs and balance of output are produced.
Economic Efficiency (EE): It is the product of TE and AE (overall efficiency).
 It can be interpreted as the potential reduction in production costs (cost

efficiency) or the potential increase in revenue (revenue efficiency) that a farm
could apply in order to operate at the point of TE and AE.
 EE enables conclusions to be drawn on whether the farm operates at optimal or

suboptimal size.

Efficiency Measuring Techniques
 Many different approaches have been applied by many researchers to measure

efficiency.
1. The parametric Stochastic Frontier Analysis (SFA) and
2. The non-parametric Data Envelopment Analysis (DEA).

The parametric approach (Stochastic Frontier Analysis (SFA):
 It is the parametric approach consists of specifying and estimating a parametric

production function frontier, and calculating technical inefficiency.
 The frontier production function is an extension of the familiar regression model

based on the microeconomic premise that a production function represents some
sort of ideal, the maximum output attainable given a set of inputs.
 The parametric approach is naturally subdivided into deterministic and stochastic

models.
 Deterministic models envelope all the observations, identifying the distance

between the observed production and the maximum production, defined by the
frontier and the available technology, as technical inefficiency.

 Stochastic approaches: Permit one to distinguish between TE and statistical
noise.
 The strengths of the stochastic frontier approach are that:
 It deals with the stochastic noise and

 It permits statistical tests of hypotheses pertaining to the structure and
 It allows the test of hypothesis concerning the goodness of fit of the model.
 Its main weakness is:
 The assumption of an explicit functional form for the technology and

 The distribution of the inefficiency terms.

The non-parametric approach (Data Envelopment Analysis (DEA)):
 It is frontier estimation approach reformulated as a mathematical programming

model using the DEA.
 DEA is defined as a mathematical programming model applied to the

observational data that provides a new way of obtaining empirical estimates of
the production functions and/or efficient production possibility surfaces that are
cornerstones of modern economics.
 This methodology was used initially to measuring the relative efficiency of a set
of Decision-Making Units (DMUs).
 The model distinguishes between pure TE and scale efficiency (SE), identifying
if increasing, decreasing or constant returns to scale are present.
 DEA is a relatively new data oriented approach for providing a relative efficiency
assessment and evaluating the performance of a set of peer entities called DMUs
which convert multiple inputs into multiple outputs.
 DEA is thus a multi-criteria decision making method for evaluating effectiveness,
efficiency and productivity of a homogenous group (DMUs).
 DEA is a linear programming methodology to measure the efficiency of multiple

DMU when the production process presents a structure of multiple inputs and
outputs.
 It is increasingly popular management tool.
 It is commonly used to evaluate the efficiency of a number of producers.
 It’s an extreme point method and it compares each producer with only the
"best" producers.
 Producer is usually referred to as a decision making unit (DMU)
 DEA has been used for both production and cost data.

Why DEA was created?
 In microeconomic production theory a firm’s input and output combinations are

depicted using a production function.
 Using such a function one can show the maximum output which can be achieved
with any possible combination of inputs, that is, one can construct a production
technology frontier.
 Some 30 years ago DEA set out to answer the question of how to use this
principle in empirical applications while overcoming the problem that for actual
firms one can never observe all the possible input-output combinations.
 Building on the ideas of Farrell (1957), the seminal work "Measuring the
efficiency of decision making units" by Charnes, Cooper & Rhodes (1978)
applies linear programming to estimate an empirical production technology
frontier for the first time.
 The main developments of DEA in the 1970s and 1980s are documented by
Seiford & Thrall (1990).
How it works?
 Utilizing the selected variables, such as unit cost and output, DEA software
searches for the points with the lowest unit cost for any given output,
connecting those points to form the efficiency frontier.
 Any company not on the frontier is considered as inefficient.
1. Utilizing some variables to establish the efficiency frontier:

 Number of employees.
 Service quality.
 Environnemental safety.
 Fuel consumption.
2. Define the efficiency frontier.

3. A numerical coefficient is given to each firm, defining it’s relative efficiency.

Graphical illustration of how DEA works
Constant return frontier

Output
e
b a
h Variable return frontier
k j
f c
o g Inputs
1. obe = Constant return frontier -- with Uniform slope

2. gbah = Variable return frointier

Measuring Technical Efficiency
 Examine the variable returns (VR’s) frontier – (gbah)
 By refering to observation C in the graph.
 TE = the ratio of input usage assuming variable returns to scale to the

actual input usage.
 TE = fj / fc
 Technical inefficiency = 1 – fj / fc or equivalently jc / fc

Scale Efficincy
 The ratio of TE measuring assuming CRs.

 SE = fk / fc
 To the TE measure assuming

 SE = fk / fj
 Scale inefficency = 1 – fk / fj = kj / fj
 If SE = 1 --- the return to scale is constant return to scale (CRs)

 If SE ≠ 1 --- the return to scale is either increasing (IRS) or decreasing return to
scale (DRS).

Mathematical formulation
Min TE ---- because we have to minimize deviation from frontier line

St
Yj ≤ Zy
TE Xj ≥ Zx
Z Ɛ R (real number)
 Where:
Yi = is a m x 1 vector of outputs by jth sample
Xj = is a n x 1 vector of inputs utilized
Y = is a k x m matrix of outputs
K = is number of household in sample
X = is a kxn matrix of inputs
Z = is a vector of weights attach to each sample HH when constructing frontier
 If TE = 1 the HH is efficienct -- it lies on line efficicncy frontier

 If TE ≤ 1 the HH is inefficnet -- it lies below the line of efficnet frontier
 TE is the ratio of input utilized by the best practice HH to the inputs utilized by
some HHj.
DEA Case
During the 1980’s the Electronic

industry is the most activity
department among the economics.
And the competition is very serious.
Hence the innovation strategies of
computer, communication and
software company are very important.
Therefore, it makes lots of sense of
evaluating the company’s or
department’s performance with the
strategy.

The Efficiency of U.S. Electronic Company
 Data
 The data is from the database in CD of
COMPUSTAT running on IBM’s PC which is
offered by Lotus.
 The first groups are the most 8 profitable companies in three departments (based
on NI from 1980-1988).
 The second groups are the least 6 profitable companies in three departments.
 The third groups are consisted with the first seven companies from group 1 and
last four from group 2.

IBM, DEC,
HWP, HIT
APPL,
AMH
CYR, TAN

IBM, DEC,
HWP, HIT BEEV, GES,
APPL, DGN, EEC,
AMH FLP, MEM
CYR, TAN

IBM, DEC, BEEV,
HWP, HIT GES,
APPL, DGN, EEC,
AMH FLP, MEM
CYR, TAN
IBM, DEC, HWP,

APPL, AMH, CYR,
TAN, BEEV, DGN,
FLP, MEM
IBM, DEC,
HWP, HIT BEEV, GES,
APPL, DGN, EEC,
AMH FLP, MEM
CYR, TAN
IBM, DEC, HWP,

APPL, AMH, CYR,
TAN, BEEV, DGN,
FLP, MEM
The Efficiency of U.S. Electronic Company
• SALES VOLUME
• EMPLOYEE
• NET INCOME
• CAPITAL
Company • CAPITAL MARGIN
• COST
& STOCK MARGIN

EFFICIENCY
SALES
PROFIT
EMPLOYEE
CAPITAL

EFFC of Computer company

EFFC of Communication company

EFFC of Software company

Conclusion
 The company whose benefit are high doesn’t mean its efficiency is high
 New company do well in the efficiency
 The performance influenced by the surrounding too.

Advantages of DEA
 No need to explicitly specify a mathematical form for the production function
 Proven to be useful in uncovering relationships that remain hidden for other

methodologies.
 Capable of handling multiple inputs and outputs.
 Capable of being used with any input-output measurement.
 Sources of inefficiency can be analyzed and quantified for every evaluated unit

Disadvantages of DEA
 Results are sensitive to the selection of inputs and outputs

=> we have to choose relevant criteria
 Best specification cannot be tested.
 The number of efficient firms on the frontier tends to increase with the number of
inputs and output variables

Chapter Six
Transportation problem

6.1. What is transportation problem?
 Transportation problem deals with the distribution of goods from several points
of supplies (sources) to a number of points of demands (destinations).
 Cost of product = production cost + distribution cost.
 Distribution cost consists of mainly the transportation cost of items from its
production (manufacturing) center to the warehouses.
 Transportation techniques are designed to minimize the distribution costs.
 The production capacity of each product in each factory is not fixed.
 The holding capacity of a warehouse or potential sales in each marketing center

is again a fixed quantity/quality which cannot be exceeded.

 Transportation problem is one of the important application of LP.
 The model is not confined to transportation problem.
 Objective: The objective is to determine the amount of commodities

which should be transported from several sources to different
destinations, at the minimum transportation cost and /or time (to
minimize the cost of transportation).
 Sources or origins: Are the places where goods originate from (like
plants, warehouses etc).
 Destinations: Are places where goods are to be shipped.

Assumptions in transportation problem
1. Total quantity of the item available at different source is equal to the required at
different stations.
2. Item can be transported continually from all sources.
3. The unit transportation cost of the item from all sources to all destination is
precisely and certainly known.
4. Transportation cost on the given rout is directly proportional to the number of

units shipped to that rout.
5. The objective is to minimize total transportation cost for the organization as a

whole.
6. All goods are homogenous, so that any origin is capable of supplying to any
destination.

General Transportation Problem Model
 A transportation problem model, which has ‘m’ sending locations (origins) and
‘n’ receiving locations (destinations), provides a framework for presenting all
relevant data.
 These are:
 Quantity supply of each origin (SSi)
 Quantity demand of each destination (DDi)
 Unit transportation cost from each origin to each destination (Cij)

Destination
D1 D2 …… Dn Total SS
To
From
S1 X11 c11 …… X1n c1n SS1
S2 X21 c21 …… X2n c2n SS2
Source . . . . . .
or
. . . . . .
Origin
. . . . . .
Sm Xm1 cmi Xm2 cm2 …… Xmn cmn SSm
∑SS
……
Total DD DD1 DD2 DDm ∑DD

Where:
 SSi is total quantity of commodity available at origin i (total supply of origin i).
 DDi is total quantity of commodities needed at destinations j (total demand of

destination).
 Cij measures (represents) the costs of shipping (transporting) one unit of

commodity from source i to destination j.
 Xij is the quantity of commodities transported from ith origin to jth destination.

 Minimize (total transportation cost) Z = C11X11 + C12X12 + C13X13 +….+ CmnXmn
Subject to:
Capacity constraints (SS constraints)

X11+X12+----+X1n= SS1
X21+X22+----+X2n= SS2
. . . .
. . . .
Xm1+Xm2+.---+Xmn= SSm
Requirements constraints (DD constraint)

X11+X21+----+Xm1= SS1
X21+X22+----+Xm2= SS2
. . . .
. . . .
Xn1+Xn2+.---+Xnm= SSn
Xij ≥ 0 for all i and j

 There are three pertinent cases in transportation problems.
 These are:
a. Balanced case: A case where total supply equals total demand (DD = SS).

b. Unbalanced case: A case where total supply exceeds total demand (SS > DD),
and to change this unbalanced case into its balanced form we should introduce
dummy receiver.
 These shipments indicate unused supply capacity.

 We can convert this unbalanced case into balanced as:

c. Unbalanced case: A case where total demand exceeds total supply (SS <
DD), and to change this unbalanced case into its balanced form we should introduce
dummy supplier.
 These shows unmet demand.

 We can convert this unbalanced case into balanced as:
Note: The transportation algorithm is based on the assumption that the model is
balanced, meaning that the DD = SS.
 If the model is unbalanced, we can always add a dummy source or a
dummy destination to restore balance as shown on (b) and (c) above.

Steps of the Transportation Algorithm
 The steps of the transportation algorithm are exact parallels of the simplex
algorithm.
Step 1: Determine a starting basic feasible solution, and go to step 2.
Step 2: Use the optimality condition of the simplex method to determine the
entering variable from among all the non basic variables.
 If the optimality condition is satisfied, stop. Otherwise, go to step 3.
Step 3: Use the feasibility condition of the simplex method to determine the leaving
variable from among all the current basic variables, and find the new basic
solution.
 Return to step 2.
 There are 3 methods to find the initial feasible solution.
 North-West Corner Method (NWCM)
 Least Cost Method (LCM)
 Vogel’s Approximation Method (VAM)
137
North West Corner method (NWC)
 The method starts at the Upper Left-hand (northwest) corner cell (route) of the tableau
(variable X11).
 Steps of solving transportation problem using (NWC).
1. Start with the cell at the upper left hand corner, and allocate as much as possible to that
cell.
2.1. If d1 > s1, move down vertically to the second row and make the second allocation of the
magnitude X21= min (s2, d1-X11) in cell (2, 1).
2.2. If d1< s1, move right horizontally to the second column and make the second allocation of
the magnitude X12= min (s1-X11, d2) in cell (1, 2).
2.3. If d1= s1, there is a tie for the second allocation because both the row and column are
exhausted. Therefore, move diagonally to the next row and column and make the second
allocation of the magnitude X22= min (s2, d2) in cell (2, 2).
3. Repeat step 2 until all the rim requirements are satisfied.
4. Check to be sure that all rim conditions are satisfied and the numbers of occupied
cells are equal to m+n-1.
Example 1: Find the initial basic feasible solution of the following problem whose
cost and rim requirement table is given below.
Destination
Supply
Source N S E W
16 13 22 17
A 200
100 100
14 13 19 15
B 350
40 300 10
9 20 23 10
C 150
150
0 0 0 0
Dummy 90
90
Demand 100 140 300 250 Z = 10770

 Example 2.

 In this method we do not consider the unit cost of transportation.
 The North West Corner cell X11 is chosen for allocation. The origin S1 has 70 items and
the destination W1 requires only 65 items.
 Hence it is enough to allot 65 items from S1 to W1.
 The origin S1 which is alive with 5 more items can supply to the destination to the right is
alive with 5 more items can supply to the destination to the right of W1 namely W2 whose
requirement is 42. So, we supply 5 items to W2 thereby the origin S1 is exhausted.
 W2 requires 37 items more.
 Now consider the origin S2 that has 30 items to spare.
 We allot 30 items to the cell (X22) so that the origin S2 is exhausted.
 Then, move to origin S3 and supply 7 more items to the destination W2.
 Now the requirement of the destination W2 is complete and S3 is left with 43 items and
the same can be allotted to the destination W3.
 Now the origin S3 is emptied and the requirement at the destination W3 is also complete.
 This completes the initial solution to the problem.

 The total cost of transportation by this method will be:
 (65 × 5) + (5 × 6) + (30 × 2) + (7 × 5) + (43× 4) = 325 + 30 + 60 + 35 + 172

= 622.
Least-Cost Method
 The least-cost method finds a better starting solution by concentrating on the

cheapest routes.
 The method assigns as much as possible to the cell with the smallest unit cost
(ties are broken arbitrarily).
 Next, the satisfied row or column is crossed out and the amounts of supply and
demand are adjusted accordingly.
 If both a row and a column are satisfied simultaneously, only one is crossed out,
the same as in the northwest-corner method.
 Next, look for the uncrossed-out cell with the smallest unit cost and repeat the
process until exactly one row or column is left uncrossed out.

• Example:

 Examine the rows S1, S2 and S3, 4 is the least cost element in the cell (S2, W1) and
(S2, W2) and the tie can be broken arbitrarily.
 Select (S2, W1). The origin S2 can supply 30 items to W1 and thus origin S2 is
exhausted.
 W1 still requires 35 more units. Hence, shade the row S2. Shading S2, we observe
that 5 is the least element in the cell (S1, W1) and examine the supply at S1 and
demand at W1.
 The destination W1 requires 35 items and this requirement is satisfied from S1 so

that the column W1 is shaded next.
 Next, we choose 7 as least element corresponding to the cell (S1, W2). We supply
35 units from S1 to W2. Now, only one row is left behind.
 Hence, we allow 7 items from S3 to W2 and 43 items S3 to W3.

 The cost of the allocation by the least cost method is:
(35 x 5) + (35 x 7) + (30 x 4) + (7 x 7) + (43 x 7)
= 890
 The quality of the least-cost starting solution is better than that of the north-
west corner method because it yields a smaller value of transportation cost.

Vogel Approximation Method (VAM)
 This method is based on the 'difference' associated with each row and column in
the matrix giving unit cost of transportation cij.
 This 'difference' is defined as the arithmetic difference between the smallest and
next to the smallest element in that row or column.
 This difference in a row or column indicates the minimum unit penalty incurred
in failing to make an allocation to the smallest cost cell in that row or column.
 This difference also provides a measure of proper priorities for making

allocations to the respective rows and column.

 In other words, if we take a row, we have to allocate to the cell having the least
cost and if we fail to do so, extra cost will be incurred for a wrong choice, which
is called penalty.
 The minimum penalty is given by this difference. So, the procedure repeatedly
makes the maximum feasible allocation in the smallest cost cell of the remaining
row or column, with the largest penalty.
 Once an allocation is fully made in a row or column, the particular row or column
is eliminated. Hence and allocation already made cannot be changed.
 Repeat the same procedure of finding penalty of all rows and columns, choosing
the highest penalty in a row or column and allotting as much as possible in the
least cost cell in that row or column.
 Thus we eliminate another fully allocated row or column. We repeat till all supply
and demand are exhausted.
 VAM is an improved version of the Least-Cost Method that generally, but not
always, produces better starting solutions.
 Steps involved in Vogel’s Approximation Method (VAM) are:
1. For each row and column, compute the difference between the lowest cost
element and the next lowest cost element of the row/column.
 The difference between the lowest cost and the next lowest cost for a given
row or column is said to be the penalty number for that row or column.
2. Select the row or column having the largest penalty number.

 If there is a tie it can be broken by selecting the cell where minimum cost of
transportation is found.
3. Assign (allocate) the maximum number of units to the lowest cost cell in the
corresponding row or column selected in step 2.
 And then eliminate a row or column whose supply is exhausted or
requirement is satisfied.
4. Repeat step 1 to 3 until a complete initial solution is obtained.

Vogel’s Method (1): Calculate differences
Destination
Supply diff
Source N S E W
16 13 22 17
A 200 3
14 13 19 15
B 350 1
9 20 23 10
C 150 1
0 0 0 0
Dummy 90 0
Demand 100 140 300 250
diff 9 13 19 10

Vogel’s Method (2): Select xDummyE as basic variable
Destination
Supply diff
Source N S E W
16 13 22 17
A 200 3
14 13 19 15
B 350 1
9 20 23 10
C 150 1
0 0 0 0
Dummy 90 0
90
Demand 100 140 300 250
diff 9 13 19 10

Vogel’s Method (3): Update supply, demand and differences
Destination
Supply diff
Source N S E W
16 13 22 17
A 200 3
14 13 19 15
B 350 1
9 20 23 10
C 150 1
0 0 0 0
Dummy --- ---
90
Demand 100 140 210 250
diff 5 0 3 5

Vogel’s Method (4): Select xCN as basic variable
Destination
Supply diff
Source N S E W
16 13 22 17
A 200 3
14 13 19 15
B 350 1
9 20 23 10
C 150 1
100
0 0 0 0
Dummy --- ---
90
Demand 100 140 210 250
diff 5 0 3 5

Destination
Supply diff
Source N S E W
16 13 22 17
A 200 4
14 13 19 15
B 350 2
9 20 23 10
C 50 10
100
0 0 0 0
Dummy --- ---
90
Demand --- 140 210 250
diff --- 0 3 5

Vogel’s Method (6): Select xCW as basic variable
Destination
Supply diff
Source N S E W
16 13 22 17
A 200 4
14 13 19 15
B 350 2
9 20 23 10
C 50 10
100 50
0 0 0 0
Dummy --- ---
90
Demand --- 140 210 250
diff --- 0 3 5

Destination
Supply diff
Source N S E W
16 13 22 17
A 200 4
14 13 19 15
B 350 2
9 20 23 10
C --- ---
100 50
0 0 0 0
Dummy --- ---
90
Demand --- 140 210 200
diff --- 0 3 2

Vogel’s Method (8): Select xAS as basic variable
Destination
Supply diff
Source N S E W
16 13 22 17
A 200 4
140
14 13 19 15
B 350 2
9 20 23 10
C --- ---
100 50
0 0 0 0
Dummy --- ---
90
Demand --- 140 210 200
diff --- 0 3 2

Destination
Supply diff
Source N S E W
16 13 22 17
A 60 5
140
14 13 19 15
B 350 4
9 20 23 10
C --- ---
100 50
0 0 0 0
Dummy --- ---
90
Demand --- --- 210 200
diff --- --- 3 2

Vogel’s Method (10): Select xAW as basic variable
Destination
Supply diff
Source N S E W
16 13 22 17
A 60 5
140 60
14 13 19 15
B 350 4
9 20 23 10
C --- ---
100 50
0 0 0 0
Dummy --- ---
90
Demand --- --- 210 200
diff --- --- 3 2

Destination
Supply diff
Source N S E W
16 13 22 17
A --- ---
140 60
14 13 19 15
B 350 4
9 20 23 10
C --- ---
100 50
0 0 0 0
Dummy --- ---
90
Demand --- --- 210 140
diff --- ---

Vogel’s Method (12): Select xBW and xBE as basic variables
Destination
Supply diff
Source N S E W
16 13 22 17
A --- ---
140 60
14 13 19 15
B ---
210 140
9 20 23 10
C --- ---
100 50
0 0 0 0
Dummy --- ---
90
Demand --- --- --- ---
diff --- ---

Z = 10,330

Transportation Problem:
Has three phases (steps):
1. Initial basic feasible solution:
 Find the initial basic feasible solutions by different methods.
2. Degeneracy
 Test the degeneracy of the solution by checking equality between the number of
allocation and m+n-1.
 If number of allocation is less than m+n-1 there is a case of degeneracy
problem.
3. Calculate improvement index values

 After calculating check whether the improvement index have negative value. If
it has negative value the solution is not optimal solution, thus improve it again.

Methods for Testing the Optimality and Improving the Initial Solution
 Once an initial solution is identified, the next step is to check its optimality.
 An optimal solution is one in which there is no opportunity cost, that is, there
is no other set of transportation routes (allocations) that will reduce the total
opportunity cost.
 Thus, we have to evaluate each unoccupied cell (represents unused route) in the
transportation table in terms of opportunity cost.
 The purpose of the optimality test is to see if the proposed solution just
generated can be improved or not.
 The solution to be checked for optimality must be non-degenerate i.e the

number of occupied cells must be m+n-1.

 The Procedure for testing optimality is analogous (equivalent) to that of the
simplex method.
 A distinction is made between basic variables, those associated with occupied

cells & non-basic variables, those associated with the empty cells.
 For each empty cell, the effect of changing it to an occupied cell is examined.
 If any of these changes are favourable, the solution is not optimal & a new
solution must be designed.
 A favourable change means an increase in the value of the objective function in

maximization problems or a decrease in minimization problems.

 There are several methods for improving the initial solution of TP.
 Among them the two important ones are:
I. The Stepping-Stone Method (SSM)
II. The Modified Distribution (MODI) Method

I. The Stepping-Stone Method (SSM)
 The SSM is an iterative procedure that exchange one variable that is in the basis
(occupied cell or a cell with positive value), with one variable that is not in the
basis (i.e is non-basic or empty or non-occupied cell or a cell with 0 value) in
such a way that the minimal transportation cost is improved.
 It is an efficient method for determining the entering variable from available non
basic variable.

Example: Optimality Test
 Using the Loop procedure
Destination
Supply
Source D E F
6 9 16
A 150 50 200
11 10 150 7 50
B 200
16 12 10 200
C 200
Demand 150 200 250
A to F +16-7+10-9 = 10
B to D +11-6+9-10 = 4
C to D +16-6+9-10+7-10 = 6
C to E +12-10+7-10 = -1
The RHS (10, 4, 6 and -1) represent Net evaluation.
 We know that transportation is the minimization problem, thus in minimization;
optimal solution is when net benefit (Δj) is ≥ 0.
 For the current problem one of the net benefit is -1 indicates that the current basic
solution is not optimal.
 Then, we have to find the improved solution by taking -1 as entering variable.
 We use the loop that produce -1.
 -1 represent entering variable smallest is 150

 In order to maintain feasibility, examine the even numbered cells at the corners
of closed loop, and select the one that has the smallest allocation..
Destination
Supply
Source D E F
6 9 16
A 150 50 200
11 - 10 150 + 7 50
B 200
16 + 12 - 10 200
C 200
Demand 150 200 250
 From 150 (order 2 that is even) and 200 (order 4 that is even) ----- the smallest is 150.
 Then select 150 as smallest allocation.

 Subtract this smallest allocation (150) from even numbered cells (order 2 and 4
of the loop) of the closed loop and add to the odd numbered cells (order 1 and
3 of the loop).
Destination
Supply
Source D E F
6 9 16
A 150 50 200
11 10 7
B 150-150 50+ 150 200
16 12 +150 10
C 200 -150 200
Demand 150 200 250

 Then the Improved solution is:
Destination
Supply
Source D E F
6 9 16
A 150 50 200
11 10 7
B 200 200
16 12 150 10 50
C 200
Demand 150 200 250

The Modified Distribution (MODI) Method
 This method is based on the dual of the TP or it uses shadow price Ui for the
source (supply constraint) and Vj for the capacity (demand constraint).
 Ui is the relative implicit contribution (value) of an additional unit of capacity

(SS) at sources i.
 Vj is the relative implicit contribution (value) of an additional unit of requirement

(DD) at destination j.
Theorem (conditions): For balanced transportation problem,

 For occupied cell: Ui + Vj = Cij
 For empty cells: Cij - ( Ui + Vj) = ∆ij
• Ui + Vj represent the opportunity cost of shipping one unit from source i to
requirement j and
• Cij - ( Ui+Vj) is the per unit net contribution to the objective function for
opening a route from i to j which is denoted by by ∆ij.

 Steps followed in MODI method to test and improve the solution are the
followings:
Step 1: From the initial basic solution, calculate Ui and Vj for each row and column
using the occupied cells.
 To start with, assign zero for a particular Ui or Vj where there are maximum
number of allocations in a row or column. If there is a tie select arbitrarily.
 Then, compute Ui ‘s and Vj ‘s for other rows and columns, respectively, using
the relation Cij = Ui+Vj for all occupied cells (i, j).

Step 2: For unoccupied (empty) cells, calculate the net contribution by using the
relation:
 ∆ij = Cij - (Ui +Vj) for all i and j
• Compute the sum of Ui and Vj and write that at the bottom left corner of that
empty cell.
• Compute the net contribution and write that at the bottom right corner of that
empty cell.
Step 3: Examine the sign of ∆ij for empty cell
 This step gives us the opportunity criteria

• If ∆ij ≥ 0 for all empty cell, the current basic feasible solution is optimum.
• If ∆ij = 0 for some empty cell, then the current solution will remain optimum
but an alternative solution exists.
• If ∆ij ≤ 0 for one or more empty cells, the current basic feasible solution is
not optimum.
 So that an improved solution can be obtained by entering unoccupied
cell (i, j) in the basis.
Step 4: Select an empty cell having the largest –ve number of ∆ij as entering
variable (cell).
Step 5: Identify the leaving variable (cell).
 The method (criteria) of identifying the leaving variable and determining the
new solution is the same as the stepping-stone method.
 Construct the loop (closed path) for entering variable (cell)
 Assign +ve and –ve sign alternatively
 Examine the loop with minus sign and select the cell with lowest (smallest)
shipment (allocation) as leaving variable.
Step 6: Generate a new solution

 The new solution can be generated by adding the shipment (allocation)
amount in leaving cell to each plus sign cells and subtracting the amount
from each minus cell in the selected loop.
Step 7: Check for optimality of the current solution.

 If not optimal, repeat the steps from step 1 till an optimum basic feasible
solution is obtained.
 The optimum solution is obtained when for all empty cell, ∆ij ≥0
Example: Optimality Test (MODI method)
 Calculate ui, vj using cij – ui – vj = 0 or cij = ui + vj for xij basic.
(let ui = 0 for row i with the largest number of basic variables)
Destination
Supply
Source D E F
6 9 16
A 150 50 200
11 10 150 7 50
B 200
16 12 10 200
C 200
Demand 150 200 250

Example: Optimality Test
 Calculate ui, vj using cij – ui – vj = 0 or cij = ui + vj for xij basic.
(let ui = 0 for row i with the largest number of basic variables)
Destination
Supply
Source D V1=6 E V2= 9 F V3=6
6 9 16
A U1=0 150 50 200
11 10 150 7 50
B U2=1 200
16 12 10 200
C U3=4 200
Demand 150 200 250
Assign any one of the Ui’s or Vj’s the value zero. Here U1= 0.
C11 − U1 − V1 = 0, 6 − U1 − 𝑣1 = 0, let u1 = 0, v1 = 6
C12 − U1 − V2 = 0, 9 − U1 − V2 = 0, V2 = 9
C22 − U2 − V2 = 0, 10 − U2 − V2 = 0, U2 = 1
C23 − U2 − V3 = 0, 7 − U2 − V3 = 0, V3 =6
C33 − U3 − V3 = 0, 10 − U3 − V3 = 0, U3 = 4
Non basic variables
Calculate cij = cij- ui - vj for xij non-basic.
C13 - U1 - V3 = A to F = 16 – 6 – 0 =10
C21 – U2 – V1 = B to D = 11 – 6 - 1 = 4
C31 – U3 – V1 = C to D = 16 – 6 – 4 = 6
C32 – U3 – V3 = C to F = 12 – 9 - 4 = -1

 We know that transportation is the minimization problem, thus in minimization;
optimal solution is when net benefit (Δj) is ≥ 0.
 For the current problem one of the net benefit is -1 indicates that the current basic
solution is not optimal.
 Then, we have to find the improved solution by taking -1 as entering variable.
 -1 represent entering variable smallest is 150

 In order to maintain feasibility, examine the even numbered cells at the corners
of closed loop, and select the one that has the smallest allocation..
Destination
Supply
Source D E F
6 9 16
A 150 50 200
11 - 10 150 + 7 50
B 200
16 + 12 - 10 200
C 200
Demand 150 200 250
 From 150 (order 2 that is even) and 200 (order 4 that is even) ----- the smallest is 150.
 Then select 150 as smallest allocation.

 Subtract this smallest allocation (150) from even numbered cells (order 2 and 4
of the loop) of the closed loop and add to the odd numbered cells (order 1 and
3 of the loop).
Destination
Supply
Source D E F
6 9 16
A 150 50 200
11 10 7
B 150-150 50+ 150 200
16 12 +150 10
C 200 -150 200
Demand 150 200 250

 Then the Improved solution is:
Destination
Supply
Source D E F
6 9 16
A 150 50 200
11 10 7
B 200 200
16 12 150 10 50
C 200
Demand 150 200 250

The Assignment Problem
 The assignment problems are characterized by a need to pair items in one
group with items in another group in a one-to-one matching.
 There are many situations where the assignment of resources like people or
machines on to different jobs.
 The problem of finding the minimum-costly assignment of a set of tasks (I

=1,…,m) to a set of agents (j = 1,…,n)
 Each task should be performed by one agent
 Each agent should perform one task
 A cost cij associated with each assignment
 We should have m = n (if not…?)
 A special type of linear programming problem, and

 A special type of transportation problem, with s i= dj = ?
 The assignment problems are characterized by a need to pair items in one group
with items in another group in a one-to-one matching.
 The assignment problem has many applications in allocations and scheduling:
Examples:
 In assigning salesmen to different regions
 Vehicles and drivers to different routes
 Products to factories
 Jobs to machines
 Contracts to binders,
 etc.

Definitions: Given n facilities and n jobs and given the effectiveness of each facility
for each job, the problem is to assign each facility to one and only one
job so as to optimize given measure of effectiveness.
 Mathematical formulation of the assignment problem

n n
Minimize Z   Cij xij
i 1 j1
Subject to :
n
 ij
x
j 1
 1, i  1,2,..., n (one job is assigned to the i th
facility)
 xij
i 1
1, j  1,2,..., n (only one person should be assigned to the j th
job )
1, if the i th person is assigned to the jth job

x ij  
0, if not

 Comparison of transportation and assignment problem
Transportation problem Assignment problem

No. of sources and destinations need not be No. of workers and jobs need to be
equal. Hence the cost matrix is not equal. Hence cost matrix is a square
necessarily a square matrix. matrix.
Xij: indicates the quantity to be transported Xij: indicates the jth job to be
from ith origin to jth destination and take assigned to ith person and it can take
any possible value. either 1 or 0.
The row and column sum is equal to ssi The row and column sum is exactly1.
and ddj.
The problem is unbalanced if ∑ss≠∑dd. The problem is unbalanced if the cost

matrix is not a square matrix.

Prototype Problem
 Assign students to mentors

 Each assignment has a ‘mismatch’ index
 Minimize mismatches
Mentor
Supply
Student Snape McGonagall Lupin
5 2 3
Harry 1
1 4 5
Draco 1
2 4 4
Goyle 1
Demand 1 1 1

Solving the Assignment Problem
 It is a linear programming problem, so we could use regular simplex method.
 It is a transportation problem, so we could use transportation simplex method.
 However, it has a very special structure, such that it can be solved in polynomial
time.
 Many such algorithms exist, but the best known (and one of the oldest) is the
Hungarian Method.
Methods of solving assignment problems
There are four methods of solving an assignment problem.

a. Hungarian assignment method;
b. Complete enumeration method;
c. Simple method; and
d. Transportation method.
The Hungarian Method
Steps followed:
1. Subtract row minimums from each element in the row.
2. Subtract column minimums from each element in the column.
3. Cover the zeroes with as few lines as possible.
4. If the number of lines = n, then optimal solution is hidden in zeroes.
5. Otherwise, find the minimum cost that is not covered by any lines.
 Subtract it from all uncovered elements
 Add it to all elements at intersections (covered by two lines)
6. Back to step 3.

The Hungarian Method – Optimal Solution
How to identify the optimal solution:
 Make the assignments one at a time in positions that have zero elements.
 Begin with rows or columns that have only one zero.

 Cross out both the row and the column involved after each assignment is
made.
 Move on to the rows and columns that are not yet crossed out to select the next
assignment, with preference given to any such row or column that has only one
zero that is not crossed out.
 Continue until every row and every column has exactly one assignment and so has
been crossed out.

The Hungarian Method – Optimal Solution
• A method, designed specifically to handle the assignment problems in an efficient

way, which is based on the concept of opportunity cost.
• A typical balanced assignment problem involving a certain number of persons

and an equal number of jobs and with an objective function of the minimization
type, the method is applied as listed in the following steps.
Step 1: Prepare a cost matrix (a square cost matrix).
Step 2: Prepare the modified matrix as follows:
 Subtract the smallest cost element in each row of the cost table from each
element in that row. As a result, there shall be at least one zero in each row of this
new table, called the Reduced Cost Table.
 In the reduced cost table, subtract the smallest cost element in each column from
every other entry in the column. As a consequence of this action, there would be
at least one zero in each of the rows and columns of the second reduced cost
table.
Step 3: Draw the minimum number of horizontal and vertical lines (not the diagonal
ones) that are required to cover all the zero elements.
 If the number of lines drawn is equal to n (the number of rows or columns) the
solution is optimal, and proceed to step 6. If the number of lines drawn is smaller
than n, go to step 4.
Step 4: Select the smallest uncovered (by the lines) cost element. Subtract this
element from all uncovered elements including itself and add this element to each
value located at the intersection of any two lines. The cost elements through which
only one line passes remain unaltered.

Step 5: Repeat steps 3 and 4 until an optimal solution is obtained.
Step 6: Given the optimal solution, make the job assignments as indicated by the
zero elements. This is done as follows:
(a). Locate a row which contains only one zero element. Assign the job
corresponding to this element to its corresponding person. Cross out the zero’s, if
any, in the column corresponding to the element, which is indicative of the fact that
the particular job and person are no more available.
(b). Repeat (a) for each of such rows which contain only one zero.
If there is no row with only a single zero element, perform similar operation in
respect of each column containing only one zero element, crossing out the zero(s), if
any, in the in which the element lies.
c. If there is no row or column with only a single zero element left, then select a
row/column arbitrarily and choose one of the jobs (or persons) and make the
assignment. Now cross the remaining zeros in the column and row in respect of
which the assignment is made.
(d). Repeat steps (a) through (c) until all assignments are made.
(e). Determine the total cost with reference to the original cost table.

Example: 1. Hungarian Method
JOB
1 2 3 4
PERSON A 10 12 19 11
B 5 10 7 8
C 12 14 13 11
D 8 12 11 9

Check whether the assignment is balance – it is 4 x 4
JOB
1 2 3 4
PERSON A 10 12 19 11
B 5 10 7 8
C 12 14 13 11
D 8 12 11 9

Step 1: Row reduction--- subtract minimum entry from each row
JOB
1 2 3 4
PERSON A 0 2 9 1
B 0 5 2 3
C 1 3 2 0
D 0 4 3 1

Step 2: Column reduction – reduce minimum element of the column
JOB
1 2 3 4
PERSON A 0 0 7 1
B 0 3 0 3
C 1 1 0 0
D 0 2 1 1

Step 3: Check each row and column has 1 zero: which indicates one job assigned
to one person
JOB
1 2 3 4
PERSON A 0 0 7 1
B 0 3 0 3
C 1 1 0 0
D 0 2 1 1
A assigned Job 2 Person Job Total

B assigned Job 3 A 2 12
C assigned Job 4
B 3 7
D assigned Job 1 C 4 11
D 1 8
Total Minimum Cost = 12+7+11+8 = 38
Example: 2 Solve the following assignment problem for optimal solution using
HAM.
 The time taken (in minutes) by workers is presented in the table.
Step 1: Prepare a cost matrix (a square matrix).

Step 2: Prepare the modified matrix.
 Subtract the smallest cost element in each row of the cost table from each element
in that row.

• As a result, there shall be at least one zero in each row of this new
table, called the Reduced Cost Table.
 In the reduced cost table, subtract the smallest cost element in each column from
every other entry in the column.

As a consequence of this action, there would be at least one zero in each of the rows
and columns of the second Reduced Cost Table.
Step 3: Draw the minimum number of lines covering all zeros.

 As a general rule, we should first cover those rows/columns which contain
larger number of zeros.

Step 4: Since the number of lines drawn is equal to 4 (= n), the optimal solution is
obtained. The assignments are made after scanning the rows and columns for
unit zeros. Assignments made are shown with squares, as follows.

The total time associated with the optimal job assignment is:

Example: 3. using the following cost matrix, determine:
(a) Optimal job assignment.
(b) The cost of assignments.
Prepare the modified matrix: the Reduced Cost Table would be as follows.

Obtain column reductions and draw the minimum number of lines to cover all zeros.

• Since the number of lines covering all zeros is less than the number of
columns/rows, we modify the table by selecting the smallest uncovered (by the
lines) cost element and subtract it from all uncovered elements including itself and
add this element to each value located at the intersection of any two lines. The cost
elements through which only one line passes remain unaltered. Accordingly, the new
table would appear as shown in the table below. The Reduced Cost Table would be:

• The optimal assignments can be made as the least number of lines covering all
zeros in table equals.
• Considering rows and columns, the assignments can be made in the following
order:
 Select the second row. Assign mechanic B to job 4. Cross out zeros at cells (C, 4)
and (E, 4).
 Consider row 4. Assign mechanic D to job 1. Cancel the zero at cell (E, 1).
 Since there is A single zero in the 5th row, put mechanic E to job 3 and cross out
the zero at (A, 3).
 There being only a single zero left in each of the 1st and 3th rows, we assign job 2
to mechanic A and job 5 to C.
12/10/2023 by: Getachew T.(MSc.) 209
The optimal job assignment and cost associated with assignment pattern is:

Example 4.

Examples: 5

Examples: 6


OR PPT 2

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

OR PPT 2

Uploaded by

Copyright:

Available Formats

Chapter Three

Duality and Sensitivity Analysis

© 2006 Prentice Hall, Inc. B–1

 In fact either of the problems can be considered original as both originate

© 2006 Prentice Hall, Inc. B–2

 Corresponding to every LPP, there is another LPP.

 The given problem is called the primal.

 The related problem to the given problem is known as the dual.

 The dual of a normal max problem is a normal min problem.

© 2006 Prentice Hall, Inc. B–3

 The dual of a dual is the primal (original LP).

© 2006 Prentice Hall, Inc. B–4

1. The number of variables in the dual problem is equal to the number of

3. If the original problem is a maximization problem, the dual is a minimization

6. The dual of the dual problem is the original LPP itself.

© 2006 Prentice Hall, Inc. B–6

The primal is maximization The dual is minimization and vice versa

1. The dual form provides an alternative form.

© 2006 Prentice Hall, Inc. B–8

Max. Z = 5x1+ 6x2

© 2006 Prentice Hall, Inc. B–9

Find dual form

© 2006 Prentice Hall, Inc. B – 11

Primal Problem Max. Z = 2X1+ X2

© 2006 Prentice Hall, Inc. B – 12

Dual Problem Min Z* = 2y1 + 4y2 + 3y3

Max Z* = -2y1 - 4y2 - 3y3 - 0S1- 0S2- MA1-MA2

© 2006 Prentice Hall, Inc. B – 14

 For Maximization when Cj - Zi ≤ O (non-positive) = Optimal solution.

Dual - optimal iteration (iteration 2)

 One of the assumptions of a LPP is the assumption of certainty.

 The process of studying the sensitivity of the optimal solution of a LPP is

The two sensitivity analysis approaches are:

I. Trial and error approach and

II. Analytical approach

© 2006 Prentice Hall, Inc. B – 18

 Changes of the coefficients of the objective function (cj)

Decision variables can be:

 Basic (in the solution)

 Non-basic (out-of the solution)

 Changes in Cj of basic variables

 One of the important parameters of LPP is the coefficient of objective function.

 If Cj is the coefficient of non-basic variable Xj:

 If optimality is to be maintained for maximization LPP, all coefficients in the

© 2006 Prentice Hall, Inc. B – 21

© 2006 Prentice Hall, Inc. B – 22

Analysis of basic decision variables:

© 2006 Prentice Hall, Inc. B – 26

© 2006 Prentice Hall, Inc. B – 27

© 2006 Prentice Hall, Inc. B – 28

 A key assumption of linear programming is that all its functions (objective

 Although this assumption essentially holds for many practical problems, it

 Therefore, it often is necessary to deal directly with nonlinear programming

 In one general form, the nonlinear programming problem is to find:

X = (X1, X2, . . . , Xn) so as to Maximize f(x),

© 2006 Prentice Hall, Inc. B – 29

 Different algorithms are used for the different types.

 We do not have the space to survey it completely.

 However, we do present a few sample applications and then introduce some of

© 2006 Prentice Hall, Inc. B – 30

 The objective function being concave guarantees that a local maximum is a

 Similarly, the objective function being convex ensures that a local

 If there are constraints:

 𝑑𝑓/𝑑𝑥𝑗 ≤ 0 at X = X, if Xj = 0