Sunteți pe pagina 1din 2

Integer programming carries a lot of meaning when the computed optimal values are constrained to be integers, such as the

number of men or women to be assigned to a job. We have two algorithms, to facilitate integer programming, namely. 1. Branch and Bound (B&B) 2. Gomorys cutting plane method (GCP).

We will use B&B in the graphical method and the GCP method in the simplex method. The graphical method is used whenever we have two variables and the simplex method can be used whenever we have more than three variables.

The only distinguishing feature between a usual LPP and an integer programming problem (IPP) is the constraint as mentioned below. Constraint: x1, x2 >= 0 and are integers

If in an integer programming problem all (or) some of the variables are permitted to take non negative integer values then the problem is called a pure integer programming problem. If only some of the variables are constrained to be integers then it is called a mixed integer programming problem.

Common approach to solve: It is possible to ignore the integer restrictions and solve the problem as a regular LPP and non integer solution can be rounded off to the nearest integer value. However this can lead to deviation from optimality and may also violate the feasibility constraints. Therefore an integer programming problem can be solved either by a) Branch and Bound algorithm (Graphical 2 Variables) b) Gomorys Cutting Plane algorithm (Computational 2 Variables and more)

Branch and Bound Algorithm: In branch and bound algorithm the problem is first solved as an ordinary LPP and then the solution space is systematically partitioned into sub problems by deleting portions that contain no feasible solution.

Gomorys Cutting plane algorithm: First solve the problem as an ordinary LPP and then introduce additional constraints (as fractional cuts) one after the other until an integer solution is obtained. If the constraint equations are all <= type we add slack variables to the constraints to convert them into equalities. Prepare the initial simplex table and solve through the primal simplex method.

S-ar putea să vă placă și