Descending path to the solution of a pure integer program: A new approach
by Professor Santosh Kumar
Abstract: Integer programming, in general, can be computationally demanding even for relatively smaller size problems. In this talk, a method for solving a pure integer program is presented, which is entirely different to conventional approaches. The proposed approach finds a descending path from the LP optimal solution to the desired integer optimal solution. The path is steered by a characteristic equation, which is first identified and then its solution is used to reach the desired integer point. The method can be easily extended to find the kth k 1 best optimal solution. In this seminar, we present various features associated with this new approach. Computation effort by this method is compared with the branch and bound method. It is hoped that the proposed approach may be computationally superior. In addition, the proposed method has many special features unique to it. The method is still in its developing stage, and more work is required to establish its potential and computational superiority, if any.
For More Information: Dr Mark Fackrell M.Fackrell@ms.unimelb.edu.au