摘要

A new knowledge-based Artificial Fish-swarm Algorithm (AFA) with crossover operator, namely CAFAC, is proposed to combat with the blind search of the original AFA. The crossover operator is explored, and the normative and the situational knowledge in the Cultural Algorithm (CA) are utilized to direct the step size as well as direction of the AFA evolution. A total of 15 high-dimensional and multi-peak functions and two typical constrained engineering benchmark problems are employed to investigate the proposed algorithm. Simulation results demonstrate that the CAFAC can yield a superior optimization performance and has a better global search ability than the AFA in dealing with both unconstrained and constrained optimization problems.

全文