人工鱼群与遗传混合算法在无人艇路径规划中的应用

A hybrid algorithm of artificial fish swarm and genetic algorithm and its application in collision avoidance of unmanned surface vessels

为了使水环境监测无人艇在监测和采集水样时能有效躲避静态障碍物,且以最优或接近最优的路径行进,提出了一种变步长和变视野的自适应人工鱼群算法与改进遗传算法混合的策略,在人工鱼完成觅食、追尾、聚群等行为后,进行遗传算法的操作。通过指数函数型衰减函数使算法前期视野和步长很大,后期视野和步长很小,提高算法的运行效率和精确性;在基本遗传算法中加入精英选择策略和保护、淘汰算子,得到全局最优解。仿真结果表明,混合算法有效克服了单一算法容易局部收敛的缺点,收敛速度快,能有效得到最优路径,计算精度高。

In order to enable the unmanned surface vehicle (USV) for water environment monitoring to effectively avoid static obstacles while monitoring and sampling water quality,and to travel in optimal or near optimal path,we propose a hybrid strategy with variable step sizes and variable view ranges,which combines an adaptive artificial fish swarm algorithm (AFSA) with an improved genetic algorithm (GA).After the artificial fish finishes foraging,tracking and clustering operations,the GA is introduced.A decaying exponential function is chosen to enlarge the vision range of and step length in the early phase and to reduce them in the late phase.This can improve the efficiency and accuracy of the algorithm.Simultaneously,an elite selection strategy,a protection operator and an elimination operator are introduced to the basic genetic algorithm to obtain the global optimal solution.Simulation results show that the hybrid algorithm can effectively overcome the disadvantages of a single algorithm,and that is easy to fall into local convergence.The proposed algorithm has a fast convergence speed and effectively obtains the optimal path with high calculation accuracy.