Abstract:

Aiming at the defects of slow rate of convergence, low accuracy and easily falling into local optimum in  artificial fish swarm algorithm (AFSA), an improved AFSA is proposed. In order to improve the accuracy, polar coordinate coding is introduced to increase the diversity in the expression solution of single matrix solution space. The polar angle is adjusted under probability by fitness in iterative process, and the uncertainty of observational results is reduced gradually. In order to increase the rate of convergence, three kinds of actions are adjusted, the random move which impacts the search direction is removedm, the search focuses on the neighborhood of the optimal solution and the possibility of repetition search is reduced. Experimental results show that the convergence and the stability of this algorithm are better than those of AFSA, adaptive artificial fish swarm algorithm (AAFSA) and niche artificial fish swarm algorithm (NAFSA), and the effectiveness of the algorithm is validated.