The Convergence and Optimization Performance of the Simplified Artificial Fish School Algorithm
Download as PDF
Fuqing Zhao, Guoqiang Yang, , Yi Zhang, Weimin Ma, Chuck Zhang, Houbin Song
As a new evolution optimization, artificial fish school algorithm has the features of high convergence speed and good performance on solving combinatorial optimization. To improve the optimization performance of the artificial fish school algorithm, a simplified artificial fish school algorithm is proposed. In the forage process, artificial fish can directly move to the optimal position in its vision distance to fasten the search speed. In the clustering behaviors, the center of the neighbor domain can be replaced by the center of the whole fish school. In the clustering process, the optimal position is replaced by optimal position of the fish school. Therefore, the computing process can proceed only by the distance of the center and optimal position of the fish school. The distance of the neighbor distance and current fish school, the utmost value of the neighbor and the distance of optimal artificial fish and all the fish neighbor domain can be simplified to shorten the running time. Simulation results show that the SAFSA has good performance on convergence speed, running time and optimization ability to high dimension function.
Artificial fish school algorithm, convergence, optimization, neighbor structure.