Differential Evolution Algorithm Based on Classification and Cross-combination.

The differential evolution algorithm is a heuristic random search optimization algorithm based on population differences; however, due to the different natures of different optimization problems, the applicability of the algorithm is limited. This paper is based on the existing algorithms, and uses heuristics thinking to improve the algorithm for blind and random optimization problems, without pertinence of the targets. Additionally, this paper introduces a concept of classification and numerical optimization problems are divided into two categories by analysis of the problem characteristics: basic combination problems and complex transformation problems. Then, a suitable method is chosen for each optimization problem corresponding to the problem category. Simultaneously, cross-combination is integrated into multiple mutation strategies and parameters for the target population to expand the search scope, quickly and accurately converging to the best solution. Then, a differential evolution algorithm based on classification and cross-combination is proposed. Finally, the effectiveness of this method is verified using the test functions set of CEC'05.