A metaheuristic plan for the Inventory Movement Problem


We present a variant of an order-picking problem where the objective function is to minimize the number of movements performed by automated guided vehicles (AGVs) between the preparation zone and the warehouse. This problem is called the Inventory Movement Optimization Problem (IMOP). In this problem, the warehouse stores containers with multiple items of the same type that have to be brought to the processing zone in order to prepare orders composed of several types of items. The preparation or processing zone accepts usually 2 to 5 slots where we can put the boxes to prepare the orders simultaneously. In the picking order process, once the boxes are set on the slots, the system has to decide which item type should be brought from storage by an AGV. Several AGVs can be used together to accelerate this process. They bring the containers with a single item that can be transferred by a robot (or an operator) in the boxes of the processing zone according to the demand. The container then returns to the warehouse. To simplify this task, many companies group orders with similar and compatible demands in batches and process one batch after another. Once a batch is completed, all the boxes of that batch are sent to the shipping zone. In our approach, we consider that waiting for a batch to be completed leads to a waste of time and significant improvements could be made if handled differently. In the IMOP, as soon as a box is filled, it is sent to the shipping zone, releasing an empty slot for a new order. To solve the IMOP, we need to find a sequence in which the orders should be processed and then another sequence in which the single item containers should be brought from the processing zone by AGVs. The objective is to minimize the number of movements of the AGVs. At the present moment, heuristics, lower bounds and a mathematical model are already helping to solve the IMOP. The presentation will focus on a new metaheuristic based on a two-step procedure. The presentation will show the general idea behind the new solving method.

EU/MEeting 2019
Eduardo García Pardo
Eduardo García Pardo
Associate Professor

Miembro fundador del grupo de investigación GRAFO, cuya línea de investigación principal es el desarrollo de algoritmos para abordar problemas de optimización, temática sobre la que versa la Tesis Doctoral del investigador y en la que se enmarcan sus publicaciones más destacadas.