Addressing premature convergence with distance based parameter adaptation in SHADE

In this paper, an analysis of a distance based parameter adaptation in Success-History based Differential Evolution (SHADE) is presented in order to show that it can have a beneficial effect on the premature convergence of the algorithm. The premature convergence of SHADE is an issue mainly in higher dimensional decision spaces. Therefore, the tests are done on the basis of CEC2015 benchmark in 10D, 30D and 50D. The results are provided and discussed. © 2018 IEEE.Ministry of Education, Youth and Sports of the Czech Republic within the National Sustainability Programme Project [LO1303 (MSMT-7778/2014)]; European Regional Development Fund under the Project CEBIA-Tech [CZ.1.05/2.1.00/03.0089]; Internal Grant Agency of Tomas Bata University [IGA/CebiaTech/2018/003]; COST (European Cooperation in Science Technology) [Action CA15140]; Improving Applicability of Nature-Inspired Optimisation by Joining Theory and Practice (ImAppNIO); High-Performance Modelling and Simulation for Big Data Applications (cHiPSet) [Action IC1406]; A.I. Lab at the Faculty of Applied Informatics, Tomas Bata University in Zli