A New Parallel Genetic Algorithm Based on TriBA Topological Structure

Abstract

In order to advance the speed of solving a large combinatorial problem, this paper proposes a new master-slave parallel genetic algorithm (PGA) based on the triplet based architecture (TriBA) topological structure. With the special topological structure by which the problem can be mapped into this model, the TriBA is able to realize the parallel computing in child topological structures and the parallel real-time communication. The theoretical analysis proves that the proposed TriBA PGA can enhance the computation speed and decrease the communication costs, thereby resulting in a novel PGA model to handle the large combinatorial problems.