Professional Documents
Culture Documents
(3rd sem.)
IPP
Applicati Submitted By
on in
Finance
IPP
Applicati
on in
Finance
IPP
Applicati
on in Group No:2
Finance
IPP
Applicati
on in
Kundan
Finance Kumar Shaw (05)
Tamanna
Submitted To: Dr. Arindam Das
Parvin (13) MBA (General)
Kazi Abu
Centre for Management studies
The University of Burdwan
Zafar Md. Yasin (23)
Concept and Definition
Integer programming problem is a special class of LPP where all or some variable are constrained to assume
non negative integer values. In linear programming each of the decision variables as well as slack variables is
allowed to take any real or fractional value. Often this is a realistic assumption. For instance, we might easily
produce 102⅔gallons of a divisible good such as wine. It also might be reasonable to accept a solution giving
an hourly production of automobiles at 58⅖ if the model were based upon average hourly production,
and the production had the interpretation of production rates. At other times, however, fractional solutions
are not realistic, and we must consider the optimization problem.
n
∑ CjXj
Maximize Z = j =1
n
∑ aijxj =bi
Subject to: j=1 (I = 1,2,…m)
X j integer (for some or all j = 1, 2, . . . , n).
Classification :
Linear integer programming problem can be classified into three parts:
1. Pure Integer Problem: in a problem when all variables are required to take integral values then it is
known as pure (all) integer programming problem.
2. Mixed Integer Problem: In a problem when some of the variables are restricted to integral values,
whereas others can assume any real value, the problem is called a mixed integer programming
3. 0-1 Programming problem: In a problem if all the variables are allowed to take values of either 0 or 1,
then the problem is called zero one programming problem. It is also called Standard Discrete
Programming problem
The general idea of the method is first to solve the problem as a continuous Model (Linear program).
¿
Suppose that Xr is an integer-constrained variable whose optimum continuous value
X r is
fractional. The range
¿
[ X r ] <xr < [Xr] +1
Cannot include any feasible integer solution. Consequently, a feasible integer value of xr must satisfy
one of two conditions
¿ ¿
Xr ≤ [ X r ] or Xr ≥ [ X r ] +1
These two conditions when applied to the continuous model result in two mutually exclusive problems
¿ ¿
X X
created by imposing the constraints Xr ≤ [ r ] and Xr ≥ [ r ] +1 on the original solution space. In this
case it is said that the original problem is branched (or partitioned) into two sub-problems.
The branch and bound method can summarize in the following steps:
n
∑ a ijxj=bi,
Subject to j =1 i = 1,2,…,m
Relaxed the integer restriction on the variables and obtain the optimal solution of the following
associated relaxed LP problem
n
∑ Cj Xj
Max Z = j=1
n
∑ a ijxj=bi ,
Subject to j =1 i = 1,2,…,m (LP-1)
And Xj ≥ 0 j=1,2,…,n
Here it may be noted that if solution to Lp-1 is infeasible or unbounded then solution to be given All Integer
Programming Problem is also infeasible or unbounded respectively.
If the optimal solution to LP-1 is feasible for AIPP then it is an optimal solution to AIPP also. Otherwise go to
step 2. Let the optimal value of the objective function of LP-1 be Z 1
Step 2: Branching Step
(a) The value Z1 will be considered as upper bound on objective value for integer LP Problem. Let it be
denoted by Zu. the lower bound on integer LP problem can be obtain by rounding off to integer all values
of the variables.
(b) Let F denotes the feasible solution set for LP-1 and from it select a non-integer component
Say Xk, let Xk denote the non integer value of Xk.
(c) Branch or partition, the problem LP1 into two new linear programs (called nodes) by adding two mutually
exclusive constraints.
¿ ¿
xk ≤ [ X k ] and xk ≥ [ X k ] + 1
X¿ X¿
To the original LP problem, where [ k ] is the integer portion of the current k of the variable Xk.
obviously it is done to exclude the non-integer value of the chosen component, i.e., X k. The two new linear
programs are as follows:
n n
∑ Cj Xj ∑ Cj Xj
(LP-II) Max Z = j=1 (LP-III) Max Z = j=1
n n
∑ a ijxj=bi , ∑ a ijxj=bi ,
Subject to j =1 Subject to j =1
¿ ¿
And Xk ≤ [ X k ] And X
Xk ≤ [ k ] + 1
Xj ≥ 0 Xj ≥ 0
Step 3:
Bound Step Obtain optimal solution of LP II and LP-III. Let the optimal value of the objective function of
LP-II be Z2 and that of LP-III be Z3.These two values establish on the IP objective value over each of the sub
problems LP-II and LP-III. Let the lower bound be denoted by ZL.
Step 4: Fathoming step Examine solution of both LP-II and LP-III, which might contain optimal
point. Exclude a sub problem from futher consideration if:
(a) Any of the sub problem has a infeasible solution it is said to be fathomed by infeasibility or
(b) ZL ≥ Zu. It said to be fathom by bounding. Or
(c) ZL is obtained at an integer feasible solution in any sub-problem and Z L < Zu. It is said to be
fathomed by integrality.
(d) If in any of the sub-problem integer value variables do not take integral values, but the solution is better
than that of current lower bound, then this value is considered as upper bound for the sub-problem in
question and come back to step 2.
A node is fathomed when we can deduce that there is nothing to be gained by examining it further.
Step 5:Termination The procedure of branching and bounding continues until no further sub-
problems remains to be examined. At this stage the integer solution corresponding to the current lower
bound is the optimal all- integer programming solution.
To solve an integer programming problem (IPP) using the cutting plane algorithm, we first drop the
integer requirements and obtain the linear programming (LP) relaxation is solved using the simplex method in
the usual way. If all the variables in the optimum solution assume integer values, we have found an optimal
solution to the IPP. Otherwise, we introduce a cut using a constraint in the optimal solution which has a
fractional right-hand-side (bi) value. The cut is then introduced into the optimal solution to LP relaxation and
the revised problem is solved using simplex method, with some modification. Cuts are introduced until all the
variables assume integer values.
Example
There are four projects under consideration. Assume that the project run into three years. Total available
funds are Rs 75000(to be used at the rate of Rs25000/- each year). The expected profit and cost break-up is as
follows
cost
Mathematical formulation
Let xj be the amount invested in project j(j=1,2,3,4). then the appropriate mathematical formulation of the
given problem as a zero-one integer programming problem is
Subject to constraints :
xj≤ 1, xj ≥ 0 and are integer (j =1, 2, 3 ,4 ). Restriction on variables indicates that xj must be 0
or1
Investment problem
Frequently, companies face the situation of selecting one or more R & D projects for other
investment opportunities among several competing projects. The ones accepted would get the value
1 and though not accepted assigned 0.
Example:
A company is considering 5 investments. The net present value (NPV) and the initial cash outflow
required for each of them is given below:
1 Rs.20000 Rs.12000
2 28000 14000
3 18500 7000
4 27500 13000
5 31000 16000
(1) The company should not make more than three investments.
(3) if the company invest in investment 4 then it should not invest in invest 5.
Formulate this as an IPP, taking the objective function as maximization of NPV.
Solution:
Let x1, x2, x3, x4 & x5 be the variables indicating investment in the investments 1,2,3,4 & 5,
respectively. Each of the variables can take a value of 0 (if an investment is not made) or 1 (if the
investment is made). Thus, we define
xi (i= 1,2,3,4,5) = 0 if investment is not made,1 otherwise
Constraints: Since the cash availability is rs 48000, the constraint may be expressed as
12000x1+14000x2+7000x3+13000x4+16000x5≤ 48,000
Requirement (2) states that if x3=1, then x2must also equal 1. now if we add the constraint x3≤ x2 or
Finally, the requirement that investment in both investment 4 & 5 can not be made may be taken
care of by adding the condition x4+x5 ≤ 1. thus if x4 = 1 , x5 would have to be zero, add if x4= 0, then
x5 may be 0 or 1
Maximize Z= 20000x1+28000x2+18500x3+27500x4+31000x5
Subject to
X1 + X2 + X3 + X4 + X5 ≤ 3
- X2 + X3 ≤0
X4 + X5 ≤ 1
X 1, X 2, X 3, X 4 & X 5 = 0, 1.
Application in portfolio Management
A project manager in a company is considering a portfolio of 10 large project investments. These
investments differ in the estimated long-run profit (net present value) they will generate as well as in
the amount of capital required.
Let Pj and Cj denote the estimated profit and capital required (both given in units of millions of £) for
investment opportunity j (j=1,...,10) respectively. The total amount of capital available for these
investments is Q (in units of millions of £)
Investment opportunities 3 and 4 are mutually exclusive and so are 5 and 6. Furthermore, neither 5
nor 6 can be undertaken unless either 3 or 4 is undertaken. At least two and at most four investment
opportunities have to be undertaken from the set {1,2,7,8,9,10}.
The project manager wishes to select the combination of capital investments that will maximise the
total estimated long-run profit subject to the restrictions described above.
Formulate this problem using an integer programming model and comment on the difficulties of
solving this model. (Do not actually solve it).
What are the advantages and disadvantages of using this model for portfolio selection?
Solution
Variables
Constraints
SUM{j=1,...,10}Cjxj <= Q
x5 <= x3+ x4
x6 <= x3+ x4
at least two and at most four investment opportunities have to be undertaken from the set
{1,2,7,8,9,10}
x1 + x2 + x7 + x8 + x9 + x10 >= 2
x1 + x2 + x7 + x8 + x9 + x10 <= 4
Objective
maximise SUM{j=1,...,10}Pjxj
The model given above is a very small zero-one integer programming problem with just 10 variables
and 7 constraints and should be very easy to solve. For example even by complete (total)
enumeration there are just 210 = 1024 possible solutions to be examined.
The advantages and disadvantages of using this model for portfolio selection are:
can be easily extended to deal with a larger number of potential investment opportunities
can be used for sensitivity analysis, for example to see how sensitive our portfolio selection decision
is to changes in the data
the model fails to take into account any statistical uncertainly (risk) in the data, it is a completely
deterministic model, for example project j might have a (known or estimated) statistical distribution
for its profit Pj and so we might need a model that takes this distribution into account
An IP solution can never be better than the solution to the same LP problem. The integer problem is
usually worse in terms of higher cost or lower profit.
Although enumeration is feasible for some small integer programming problems, it can be difficult or
impossible for large ones.