Continuous Function Optimization through Perturbation of Adaptive Differential Evolution with Optional External Archive
Michiharu Maeda, Ryu Aishima · 2024
This article presents continuous function optimization through perturbation of adaptive differential evolution with optional external archive. Adaptive differential evolution with optional external archive implements a mutation strategy of DE/current-to-pbest including optional external archive and controls the scale factor and the crossover rate in an adaptive manner. Perturbation has a physical phenomenon that the motion by the contribution of the main force is disturbed by the influence subsidiary power. By introducing perturbation, our algorithm tends to be able to explore a solution efficiently and to avoid to local optima. In order to show the validity of our algorithm, numerical experiments are examined in comparison with existing algorithms.