A Differential Evolution Algorithm with Q-Learning for Solving Engineering Design Problems

Abstract

In this paper a differential evolution algorithm with Q-Learning (DE-QL) for solving engineering Design Problems (EDPs) is presented. As well known the performance of a DE algorithm depends on the mutation strategy and its control parameters namely crossover and mutation rates. For this reason the proposed DE-QL generates the trial population by using the QL method in such a way that the QL guides the selection of the mutation strategy amongst four distinct strategies as well as crossover and mutation rates from the Q table. The DE-QL algorithm is well equipped with the epsilon constraint handling method to balance the search between feasible regions and infeasible regions during the evolutionary process. Furthermore a new mutation operator namely DE/Best to current/1 is proposed in the DE-QL algorithm. In this paper 57 EDPs provided in Problem Definitions and Evaluation Criteria for the CEC 2020 Competition and Special Session on A Test-suite of Non-Convex Constrained Optimization Problems from the Real-World and Some Baseline Results are tested by the DE-QL. We provide our results in Appendixes and will be evaluated with other competitors in the competition.