基于(1+1)进化策略的港口AGV路径规划改进进化算法

Improved evolution algorithm for port AGV path planning based on(1+1)evolutionary strategy

针对基于遗传算法、强化学习等优化算法在港口AGV路径规划中存在容易陷入局部最优和收敛速度较慢等问题,提出一种基于(1+1)进化策略的港口AGV路径规划改进优化算法,并结合港口AGV路径规划问题特点,对算法中的适应度函数从路径长度、路径平滑度和碰撞风险3个方面进行改进,构建适用于港口AGV路径规划的适应度指标。经港口AGV路径规划仿真实验验证,(1+1)-ES算法的优化性能较差分进化算法(DE)和遗传算法(GA)优势明显,能够提高港口AGV路径规划中的路径搜索效率。

Aiming at the problem that optimization algorithms based on genetic algorithm and reinforcement learning are prone to fall into local optimal and slow convergence rate in port AGV path planning,an improved optimization algorithm based on(1+1)evolutionary strategy for port AGV path planning was proposed.Combined with the characteristics of port AGV path planning,the fitness function of the algorithm was improved from three aspects:path length,path smoothness and collision risk,and the fitness index suitable for port AGV path planning was constructed.The simulation experiment of port AGV path planning shows that the optimization performance of(1+1)-ES algorithm has obvious advantages over differential evolution algorithm(DE)and genetic algorithm(GA),which can improve the path search efficiency of port AGV path planning.