“货到人”拣选系统的两阶段随机规划模型

A Two-Stage Stochastic Programming Model for the “Goods to People” Picking System

摘要: 在“货到人”智能仓储拣选体系中，机器人的合理调度和任务分配影响着系统的效率与成本。为此本文同时考虑“货到人”拣选系统中的机器人调度和任务分配，建立了机器人空闲时间不确定的两阶段随机规划模型。第一阶段，对于给定的机器人数量，作出是否调度机器人的上层策略，使得机器人完成所有任务耗费的总期望成本最小；第二阶段，对于某个场景，作出如何合理地进行任务分配的下层策略，使得机器人完成所有任务的空闲时间成本最小。然后利用遗传算法对此模型进行求解，并通过实例仿真验证了模型的可行性。

Abstract: In the intelligent warehouse picking system of “goods to people”, the reasonable scheduling and task assignment of robots affect the efficiency and cost of the system. Therefore, this paper considers both robot scheduling and task assignment in the “goods to people” picking system, and establishes a two-stage stochastic programming model with uncertain robot idle time. In the first stage, for a given number of robots, the upper-level strategy of whether to schedule robots is made to minimize the total expected cost of robots completing all tasks; In the second stage, for a certain scene, the lower-level strategy of how to allocate tasks reasonably is made to minimize the cost of idle time for the robot to complete all tasks. Then the genetic algorithm is used to solve the model, and the feasibility of the model is verified by an example simulation.