Success-history based parameter adaptation for Differential Evolution

TLDR

Differential Evolution is a simple yet effective numerical optimization method, but its performance heavily depends on control parameter settings. We propose a success‑history based parameter adaptation technique for DE that uses a historical memory of successful settings to guide future parameter selection. The method is evaluated on 28 CEC2013 problems, CEC2005 benchmarks, and 13 classical benchmarks, comparing its performance to other DE variants. Results show that DE with this adaptation is competitive with state‑of‑the‑art DE algorithms.

Abstract

Differential Evolution is a simple, but effective approach for numerical optimization. Since the search efficiency of DE depends significantly on its control parameter settings, there has been much recent work on developing self-adaptive mechanisms for DE. We propose a new, parameter adaptation technique for DE which uses a historical memory of successful control parameter settings to guide the selection of future control parameter values. The proposed method is evaluated by comparison on 28 problems from the CEC2013 benchmark set, as well as CEC2005 benchmarks and the set of 13 classical benchmark problems. The experimental results show that a DE using our success-history based parameter adaptation method is competitive with the state-of-the-art DE algorithms.