基于冯•诺依曼邻域结构的人工鱼群算法

提出了一种基于冯¢ 诺依曼邻域结构的人工鱼群算法. 每条人工鱼只和与自己相连的上下左右的人工鱼进行信息交换, 从而减少了计算邻域中心位置和极值位置的计算量, 有效地维持了种群的多样性, 加快了算法的运行速度. 在觅食行为中, 人工鱼通过直接移动到搜索到的较好位置, 来加快搜索速度. 在随机游动行为中, 人工鱼以小半径进行搜索, 因此算法的优化精度得到了提高. 采用动态调整人工鱼视野和步长的方法, 较好地平衡了全局搜索能力和局部搜索能力. 仿真和实例计算结果表明, 该算法具有更好的优化性能.

Artificial fish-swarm algorithm based on Von Neuman neighborhood

An improved artificial fish-swarm algorithm based on Von Neuman neighborhood is proposed. In the algorithm each artificial fish is assumed to exchange messages only with neighboring artificial fish. This assumption reduces the computation time in finding the center and the extremum location within the neighborhood, while effectively retains the variety of the fish-swarm and increases the running speed of the algorithm. In the behavior of preying, the artificial fish will move directly to the superior position, raising the speed of searching. In the behavior of random swimming, the artificial fish will search the object in a region of small radius, improving the accuracy of searching. By dynamically adjusting the visual field and the step of searching for artificial fish, a compromise can be made between the ability of global search and the ability of local search. The experimental results show that the proposed algorithm has better optimization performance.