Hardware architecture and FPGA implementation of a parallel elitism-based compact genetic algorihm

This paper presents hardware architecture and FPGA implementation of a parallel genetic algorithm called elitism-based compact genetic algorithm (EC-CGA). EC-CGA uses the probability model migration instead of migrating individuals in the population. The performance of the EC-CGA is enhanced by elitism-based approach. We propose scalable hardware architecture for the EC-CGA. The highly parallelized hardware units support scalable number of variables and precision. The benchmark problems and experimental results show the significant performance improvement.