Differential evolution algorithm based on self-adaptive adjustment mechanism

Differential evolution algorithm is a strong effective method for optimization problems. Parameter setting is one crucial point to improve the DE´s performance. Hence, a DE based on self-adaptive adjustment mechanism (SAMDE) is proposed to tune the size of offspring population NP besides mutation scale factor F and crossover constant Cr automatically. Moreover, the proposed algorithm applies a DE strategies pool to adjust mutation strategy during different evolution stage. Testing the algorithms on multimodal or complex continuous benchmark functions, we find that the proposed SAMDE performs better than classic DE algorithms. Performance comparisons with JADE are also significant.