Asynchronous differential evolution with convex mutation

Abstract

Asynchronous differential evolution (ADE) is recently introduced variant of differential evolution (DE). In ADE the mutation, crossover, and selection operations are performed asynchronously whereas in DE these operations are performed synchronously. This asynchronous process helps in good exploration and well suited for parallel optimization. In this study the strength of ADE is enhanced by incorporating convex mutation. Convex mutation efficiently utilizes the information of the parents which assists in faster convergence. The proposal is named ADE–CM. The potential of the proposal is evaluated and compared with state-of-the-art algorithms over a selected noisy benchmark functions consulted from the literature.