Robot dispatching problem for compact storage and retrieval systems with dual-command transactions

Abstract

A robotic compact storage and retrieval system (RCSRS) is a grid-based system that has been popularly implemented by online retailers around the world recently. This article addresses a robot dispatching problem for RCSRS, aiming to reduce robot empty travelling time and workstation idle time. The robot dispatching problem is to assign tasks to robots when the task sequences in workstations are given. Additionally, the robots follow a fast dual-command transaction process in this problem. Hence, a mixed integer programming model catering for the dual-command transaction process is formulated to optimize system performance. Furthermore, an adaptive neighbourhood search with an ε-greedy algorithm is proposed to solve the problem. The experimental results substantiate the superiority of the authors' proposed algorithm compared to various algorithms used in practice and discussed in the relevant literature. Moreover, optimal system configurations for the instances generated are suggested.