60 / 2024-04-18 10:04:28

Modelling, analysis, and control for the Robotic Order Consolidation Process

Logistics; Robotic warehousing system; Queuing network model; Markov decision process; dynamic allocation policy

The integration of robotic technology within intralogistics is an area that has seen consistent growth. This technology offers the advantage of flexibility in meeting varying demand requirements, operates without interruption, and can efficiently work in tandem with human labour or other robotic systems. This paper focuses on a specific application of robotics in order consolidation, which involves the collaboration of two distinct types of autonomous mobile robots (AMRs) to streamline the fulfilment process. The first type of robot, known as sorting robots, is responsible for sorting items and consolidating them to form orders as per customer specifications. The second type, known as switching robots, automates the process of removing containers with fulfilled orders and replacing them with empty ones in preparation for subsequent orders. In order to examine the interactions between the two processes, we developed a new analytical model to evaluate performance metrics. We model the sorting and switching processes as two separate closed queuing network (CQN) models, using batch service queues to capture the process of items forming customer orders. A machine repairman model is developed to integrate the above CQN models and batch service queues and obtain the throughput capacity of the whole process. Based on the model, we analyse and evaluate the throughput capacity of several policies, including zoning and non-zoning policies for the sorting process, as well as the first-come, first-switch policy and priority policy for the switching process. In addition, we developed a dynamic robot allocation policy with the objective of reducing the long-term average cost of the robotic order consolidation process. We verify the performance of the integrated queuing model through simulation. Experimental studies are carried out to analyze the efficiency of the integrated queuing network and the dynamic robot allocation policy.