An improved GAFSA with adaptive step chaotic search

Abstract

In order to overcome the drawbacks of Global Artificial Fish Swarm Algorithm (GAFSA), such as slow convergence, low precision, difficult to give the initial step, a lot of invalid calculation and so on, a modified GAFSA (ADP_CS_GAFSA) is proposed. According to the convergence condition, ADP_CS_GAFSA can adjust the step length and other parameters automatically to improve the performance of the algorithm. The adaptive chaos search is also used to improve the optimization accuracy. The strategy of randomly search in large scale and chaotic search in small scale is also used. When the convergence turned to the optimal value, the convergence rate becomes low, thus some condition is meet, the step of GAFSA will be expanded or shrank, and the process repeats until the step down to the set value. The computing results of some international standard test functions show that the accuracy and the convergence speed of this method is improved indeed.