CWDE: A Novel LSHADE Variant with Cauchy Distribution-based Weight Selection
Kaiyu Wang, Sicheng Liu, Lingyu Qi, Jiaru Yang, Shangce Gao · 2023
Differential evolution (DE) is a long-standing methodology for resolving intricate optimization concerns. LSHADE is an effective variation of DE. It has been successful in numerous applications and is highly regarded in the field. Our paper introduces a novel variant of LSHADE, namely CWDE, which replaces the conventional greedy selection in DE with a weight selection method grounded on Cauchy distribution (CW). To evaluate CWDE's performance, we test it on the 2017 IEEE Congress on Evolutionary Computation (CEC) benchmark functions. The experimental results confirm that CWDE outperforms LSHADE and other advanced competitors.