GAP/D: VLSI Hardware for Parallel and Adaptive Distributed Genetic Algorithms
Kazutaka Kobayashi, Norihiko Yoshida, Shuji Narazaki · 2009
This paper presents GAP/D, a VLSI implementation of a dynamic adaptation scheme for the frequency of inter-deme migration in distributed genetic algorithms (GA). Distributed GA, or multi-deme-based GA, uses multiple populations which evolve concurrently. The purpose of dynamic adaptation is to improve convergence performance so as to obtain better solutions. Through simulation experiments, we proved that our scheme achieves better performance than fixed frequency migration schemes.