A Genetic Algorithm based Approach for Topological Optimization of Interconnection Networks

The paper addresses the two terminal reliability while designing the interconnection networks. Thus a topological optimization problem is defined as the existence of at least a reliable path between a pair of nodes satisfying the predefined cost of the network. A new method based on Genetic Algorithm is proposed to solve the above said problem. In the proposed method the chromosome as well as the genes are efficiently encoded so that the cross over provides the optimal solution with better convergence rate. The reliability of some benchmark interconnection networks are evaluated by the proposed method. The population size and the computational time of the said networks as reported in this paper ensures that the proposed method converges to it's optimal solution in very few cpu secondss, while maximizing the value of the reliability of the said network to a greater extent.