Comparative study of the distance/improvement based shade

Abstract

In this paper, a comparative study of seven variants of the Success-History based Adaptive Differential Evolution (SHADE) algorithm is presented with the aim to show the influence of improvement and distance based parameter adaptation to its performance. Seven algorithms range from improvement only based SHADE through a balanced improvement and distance based DIb_SHADE to distance only based Db_SHADE. The algorithm set is compared on the CEC2015 benchmark set, results are statistically tested, commented and possible future research directions are provided. © ECMS Lars Nolle, Alexandra Burger, Christoph Tholen, Jens Werner, Jens Wellhausen