Abstract:Since the artificial fish-swarm algorithm(AFSA)and firefly algorithm(FA) are easily converging to local op-timum and have low accuracy in the optimization process for solving multi-dimensional and multi-extreme value functions, an algorithm called Fish swarm algorithm with Lévy flight and firefly behavior(LFFSA) is proposed, which introduces the migration strategy of firefly algorithm into the two behavior patterns of fish swarm as:the swarming and the preying behaviors. Furthermore,the Lévy flight is introduced into the search strategy. Besides,nonlinearity visual and step length based on dynamic parameter are simultaneously considered for limiting the search band. Simulation results demonstrate that the LFFSA has a better performance in convergence speed and optimization accuracy.

Fish swarm algorithm with Lévy flight and firefly béhavior