Simultaneous pallet poading problem: Picker-to-parts system

• Main Author: Correia, Andreia Soares
• Publication Date: 2024
• Format: Master thesis
• Language: eng
• Source: Repositórios Científicos de Acesso Aberto de Portugal (RCAAP)
• Download full: http://hdl.handle.net/10400.22/26449
• topic: Order Picking; Picker-to-parts; Autonomous Mobile Robot; Mixed Palletizing; Warehouse; Automatic; Robot Picking System; Dynamic Picking; Time Slots; Pallets; Customers

This study focuses on a novel robot order picking optimisation problem. Therefore, a Decision Support System was developed to optimise the picking and delivery process in a picker-to-parts system assisted by Autonomous Mobile Robots (AMRs). This problem takes place in the context of logistics for a fictitious retail company. The main goal is to automate the order picking process to reduce the total operation time and optimise the efficiency of Autonomous Mobile Robot routes. This research has been motivated by the need to improve the efficiency of operations in automated warehouses. Order picking for several customers can be complex and sometimes time-consuming. Therefore, it became crucial to study the motion of the AMRs to follow a predefined delivery sequence to optimally allocate the pallets and deliver them in the right order to meet customer demands. The problem has been modelled using a Mixed Integer Programming approach, and the software IBM ILOG CPLEX Optimization Studio was used to implement it. The model uses a lexicographic objective function that optimises, in order of priority, the makespan, the number of pallets used, and the unpairing of Autonomous Mobile Robots and pallets in consecutive time slots. Additionally, a total of 81 different instances were tested, varying the sizes of the sets and the values of the parameters, to evaluate the performance of the model under different conditions. The results indicate that the proposed model effectively reduces operation time by optimising the movement and allocation of AMRs, especially in scenarios with moderate complexity. In cases where the number of customers and pallets increased significantly, the model still found optimal solutions within reasonable computation times, although in some instances the time limit imposed was reached before full optimisation could occur. It is concluded that the proposed model efficiently optimises robot routes and pallet allocation, showing significant improvements in order picking efficiency in automated warehouses. However, the model faced some computational challenges as the problem complexity increased, which suggests that future work could explore heuristic or metaheuristic approaches to complement the optimisation in larger-scale instances.