Improved self-adaptive bat algorithm with step-control and mutation mechanisms

Abstract The bat algorithm (BA) is a new search optimization algorithm, inspired by bats’ echolocation behavior. However, it is prone to fall into local optima and has low solution accuracy. This study proposes an improved self-adaptive bat algorithm (SABA) with adaptive step-control and mutation mechanisms. The step-control mechanism uses two frequencies to adapt the step sizes used for the global and local searches, and the mutation mechanism could improve the algorithm’s ability to avoid local optima. SABA’s parameters are analyzed to ensure convergence. Its optimization and convergence performance are experimentally studied using 12 unimodal and multimodal functions; compared with a range of baseline algorithms, it can effectively avoid local optima and exhibits high solution accuracy. Further, its practical performance is evaluated using engineering optimization problems.