You are on page 1of 3

Special Cases

When learning to use network models, it is helpful to recognize several special cases of network
flow programming. These are the transportation, assignment, shortest path, and maximum flow
models. The problems differ primarily in the set of arc parameters that are relevant or the
arrangement of nodes and arcs. For example, for the transportation model only the arc costs are
relevant and all arcs originate at one set of nodes and terminate at another. We describe in this
section the classes in terms of the network model defined above and note that for each only a
subset of the parameters is relevant. All irrelevant parameters will take on the default values. For
the special network models of this section we set the parameter default values as: 0 for the lower
bound, M (a large number) for the upper bound, 0 for the cost, and 1 for the gain.

Using the default values for parameters, all of the classes can be solved using algorithms defined
for the more general case. There are, however, a number of algorithms for solving each class that
do not use the irrelevant parameters at all. In this way the special case algorithms can be more
efficient than the more general ones. Fig. 2 shows the relationships between the various network
flow programming models and linear programming. The models on the left are the least general.
As we move to the right, the problems become more general. Thus all the problems to the left of
the generalized minimum cost flow problem can be solved with an algorithm designed for the
generalized problem. The generalized problem is itself a special case of the linear program.

Figure 2. The relationships between network problems

ASSIGNMENT PROBLEM
A special case of the transportation problem is the assignment problem which occurs when each
supply is 1 and each demand is 1. In this case, the integrality implies that every supplier will be
assigned one destination and every destination will have one supplier. The costs give the charge
for assigning a supplier and destination to each other. Or as the name indicates, the problem is
also about assigning some objects to some other objects so that certain requirements are
satisfied as much as possible. If you are to assign office suites to your company employees
ranging from the CEO to a clerk, you should know a little about assignment problem and ways to
solve this problem. This case study will tell you what the problem is about, its detailed
formulations and implementations.
Consider the situation of assigning m jobs to n machines.
MACHINE Assignment LP Model:
1 2 3 .. n
n n
1 C11 C12 C13 .. C1n
2 C21 C22 C23 .. C2n min y =  CijXij
JOB

3 C31 C32 C33 .. C3n i=1 j=1

 X =1 I = 1,2,3..n
..

m Cm1 Cm2 Cm3 .. Cmn ij

consider: MACHINES
 xij=1 j=1,2,3..n
1 2 3 4

1 1 4 6 3
JOB 2 9 7 10 9

3 4 5 11 7

4 8 7 8 5
Step 1: Get the minimun element in each row then subtract to the row elements
Step 2: Get the minimum element of each column then subtract to the column elements

1 2 3 4

1 0 3 2 2
JOB 2 2 0 0 2

3 0 1 4 3

4 3 2 0 0

Step 3: Draw minimum number of lines through some of the rows and columns such that all the
zeros are croosed out.

1 2 3 4

1 0 3 2 2
JOB 2 2 0 0 2

3 0 1 4 3

4 3 2 0 0

Step 4:Select the smallest uncrossed-out element. This element is subtracted from every
uncrossed-out element and added to every element at the intersection of two lines
smallest element = 1

Therefore: job 1 to machine 1;


1 2 3 4 job 2 to machine 3;
job 3 to machine 2;
1 0 2 1 1 job 4 to machine 4.
JOB the total cost is:
2 3 0 0 2
= 1 + 10 + 5 + 5
3 0 0 3 2 = 21

4 4 2 0 0

Note: If the optimal solution was not obtained in the preceding step, the given procedure of
drawing lines should be repeated until a feasible assignment is achived.

You might also like