A mnemonic shuffled frog leaping algorithm with cooperation and mutation

Abstract

Shuffled frog leaping algorithm (SFLA) has shown its good performance in many optimization problems. This paper proposes a Mnemonic Shuffled Frog Leaping Algorithm with Cooperation and Mutation (MSFLACM), which is inspired by the competition and cooperation methods of different evolutionary computing, such as PSO, GA, and etc. In the algorithm, shuffled frog leaping algorithm and improved local search strategy, cooperation and mutation to improve accuracy and that exhibits strong robustness and high accuracy for high-dimensional continuous function optimization. A modified shuffled frog leaping algorithm (MSFLA) is investigated that improves the leaping rule by combining velocity updating equation of PSO. To improve accuracy, if the worst position in the memeplex couldn’t get a better position in the local exploration procedure of the MSFLA, the paper introduces cooperation and mutation, which prevents local optimum and updates the worst position in the memeplex. By making comparative experiments on several widely used benchmark functions, analysis results show that the performances of that improved variant are more promising than the recently developed SFLA for searching optimum value of unimodal or multimodal continuous functions.