An improved differential evolution based artificial fish swarm algorithm and its application to AGV path planning problems

AGV path planning problems play an extremely important role in navigations of AGV. Intelligence algorithms provide an effective way to solve such complicated problems. Artificial fish swarm algorithm (AFSA) is a newly proposed promising swarm intelligence optimization algorithm, yet there still exist some disadvantages of it, such as low optimization precision and convergence rate. Aiming at these defects, an improved differential evolution based artificial fish swarm algorithm (IDE-AFSA) is proposed in this paper and applied to the global path planning of AGV. Firstly, IDE-AFSA introduces the optimal positions stored in bulletin board into the preying, following and swarming behaviors of artificial fishes, which makes the population behaviors more purposeful and directional, as well as enhance the convergence speed of the proposed algorithm. Secondly, hybrid strategy is introduced, when the information on the bulletin board does not change for a certain number of iterations, operation based on differential evolution will be carried out, which helps to keep the population diversity and make proposed algorithm escape from local optima. The optimization results on the benchmark functions demonstrate that IDE-AFSA has better performance in convergence speed, optimization precision and stability compared with AFSA. Moreover, the experimental results of global path planning of AGV further verify the feasibility and validity of proposed IDE-AFSA.