A genetic algorithm for channel routing using inter-cluster mutation

Abstract

In this paper, we propose an algorithm for the channel routing problem based on genetic approach that uses a new type of mutation, called inter-cluster mutation. The performance of genetic algorithm-based channel router is improved by incorporating problem-specific knowledge into the inter-cluster mutation operators. A solution of the channel routing problem is represented by horizontal non-constraint graph (HNCG). The clusters information in the graph of the routing solution is used in the proposed inter-cluster mutation operators. The inter-cluster mutation integrates the concept of gradient descendent method, implicitly, into the genetic algorithm based channel router. We have implemented the inter-cluster mutation operators in C, and the results of the experiment show improved performance for genetic algorithm-based channel router using inter-cluster mutation. >