带时间窗的车间搬运机器人路径优化建模及算法研究

本文以车间搬运机器人为研究对象,在考虑时间窗的前提下,求解机器人进行物料配送和成品回收场景下的路径优化问题。提出一种强化学习遗传蚁群算法,首先利用扫描法求解初始搬运机器人的数量,并将子路径节点的几何中心设置为虚拟节点,利用嵌入遗传算子的蚁群算法求解连接虚拟节点的最优路径,再利用强化学习算法求解子路径的最优结果;最后将基本成本、运输成本和时间惩罚成本的加权和作为目标解,并最终求出满足约束条件的最优解。通过与基准问题求解结果对比,验证了强化学习遗传蚁群算法的优越性。

Research on Path Optimization Modeling and Algorithm of WorkshopHandling Robotwith Time Window

This paper takes the workshop handling robot as a research object, and solves the path optimization problem of pickup and delivery with the time window. This paper proposes a method of reinforcement learning genetic-ant colony hybrid algorithm(RLGA). Firstly, the number of initial handling robots is solved by scanning method, and the geometric center of sub-path nodes is set as virtual node. The ant colony algorithm embedded with genetic operator is used to solve the optimal connection virtual node. Secondly, the optimal sub-path is solved by using the algorithm of reinforcement learning.Finally, the weighted sum of the basic cost, transportation cost and time penalty cost is taken as the target solution, and the optimal solution satisfying the constraint condition is obtained. The superiority of the reinforcement learning genetic-ant colony hybrid algorithm is verified by comparing with the results of the benchmark problem.