A Dynamic Levy Flight Artificial Sheep Algorithm with Random Mutation Strategy and Elite Mutation Strategy

Abstract

In recent years, more and more swarm intelligence algorithms are proposed and used to solve optimal problems for their better applicability and optimization performance. As swarm intelligence algorithms have been applied in more and more fields, the drawback of swarm intelligence algorithms gradually is found when they are utilized to solve optimal problems. Whether the optimization performance of a swarm intelligence algorithm is excellent or not, it can’t avoid premature or local optimization perfectly. Therefore, employing novel improvement strategies in swarm intelligence algorithms is helpful for promoting optimization performance of swarm intelligence algorithms. A new optimization algorithm named Artificial sheep algorithm (ASA) is used to solve optimal problems in some different fields, and the test results obtained by various benchmark functions being run on ASA could show that Artificial sheep algorithm (ASA) is an excellent optimization algorithm. Despite all this, ASA is possibly trapped into local optimum or premature. To improve optimal performance of ASA effectively, three improvement strategies, which include random mutation strategy, elite mutation strategy and dynamic levy flight strategy, are applied on ASA. To prove that the optimal performance of the improved ASA named NHASA is superior to ASA, different sorts of benchmark functions are run on ASA and NHASA respectively and test results verify the view above. Random mutation strategy, elite mutation strategy and dynamic levy flight strategy, which can be applied on other swarm intelligence algorithms, may promote the optimization ability of other swarm intelligence algorithms.