A fast shuffled frog leaping algorithm

Because of the weaknesses of the shuffled frog leaping algorithm (SFLA) for optimizing some functions such as a low optimization precision, a slow speed, and trapping into the local optimum easily, etc., a fast shuffled frog leaping algorithm (FSFLA) is proposed. At first, each individual of subgroups learns from the group extremum and the subgroup extremum when it is updated by the update strategy. Its boundaries are controlled by the “hit-wall” method. Secondly, the speed of this algorithm is improved by means of sorting and grouping all individuals at a regular interval. Then, in order to keep most individuals and take full advantages of the useful information in the population, a small number of individuals are randomly generated. By comparing and analyzing the experimental results of several standard test functions, the high convergence precision and fast speed of the FSFLA are validated.