Iterated Greedy Local Search for the Order Picking Problem Considering Storage Location and Order Batching Decisions

Abstract: Order picking is the process of collecting products from a specific location to complete customer orders. It is the most costly activity inside of a warehouse with up to 65% of the incurred costs. On the other hand, two important and closely related problems to the order picking are, indeed, the storage location of products and the order batching, which may affect the routes performed by pickers if they are not optimized. In the first one, the decision to take into account is about where to place the items arriving at the warehouse considering that there are several and available locations with certain capacity and, in the second one, the decision is to determine how to group customer orders into batches, which must be assigned later to pickers to perform the corresponding routes. Although these three problems are commonly studied and solved independently, recent studies have shown that their integration may result in a greater improvement. In this paper, we study an integrated picking problem that considers these three subproblems simultaneously. The problem is motivated by a real-world application in a local warehouse. The aim of this integration is to obtain the best storage location of products, order batching, and picking sequence that minimize the total picking routing cost. We first present a mixed-integer linear programming model. Given the inherent computational complexity of the problem, we propose an iterated greedy local search metaheuristic that attempts to exploit some particular properties of the model. The proposed method is extensively assessed on random pseudo-real and real-world instances. The results indicate that instance sizes with up to 8197 order lines, 654 customer orders, and 4252 products, can be solved efficiently.