Shuffled frog leaping algorithm based on enhanced learning

The paper proposes shuffled frog leaping algorithm (SFLA) based on enhanced learning, which generates a virtual general centre frog that is related to the optimal frog of each memeplex. The algorithm can utilise the superior information of each memeplex, enhance the mutual learning and use the average centre of optimal frog. In the processing of evolution, the optimal frog of sub-memeplex learns from the general centre frog and the best frog of the whole memeplex; then it enhances the learning ability of the worst frog from general centre frog. On the one hand, the evolution increases the information share and exchange among each memeplex; on the other hand, it raises the convergence velocity. The experiment results show that the new approach has better convergence speed and searching global optimum, comparing with the standard SFLA, PSO and other variants.