Genetic Algorithm Particle Swarm Optimization Based Hardware Evolution Strategy

There are many problems exist in the Evolutionary Algorithm (EA) using Genetic Algorithm (GA), such as slow convergence speed, being easy to fall into the partial optimum ,etc. Particle Swarm Optimization (PSO) can accelerate the space searching and reduce the number of convergences and iterations. The proposed characteristics of Genetic Algorithm Particle Swarm Optimization (GAPSO) are proved by many examples, when the GA, PSO and GAPSO are adopted under the same conditions, GAPSO can get the least iteration numbers and the highest evolvable success rate. It also can reduce the number of convergence iteration and raise the accuracy of searching. And the performance of PSO is inferior to the performance of GAPSO, while the GA has the worst searching performance. It also can be found that the number of initializing particles will affect the number of convergences and iterations. The larger the number of the initializing particles is, the less the number of iterations will be. Key-Words: Hardware Evolution, Evolutionary Algorithm, GAPSO, Fault Self-repair