International Conference on Operations Research - OR 2023
29.08. - 01.09.2023 Hamburg
Konferenzvortrag
M. Sc. Regina Thelen
A fix-and-optimize approach for the integrative zoning and item-to-zone assignment in pick-and-pass order picking systems
Ralf Gössinger, Regina Thelen
Order picking is the logistic warehouse process of consolidating stored items according to customer orders. Pick-and-pass systems are zoned systems in which each picker only works in one storage zone and each zone only comprises a sub-set of items. Hence, order-related containers are routed along a sequence of zones to be filled up with the items needed for order fulfillment.
When designing a pick-and-pass system, two organizational questions need to be answered: How should the zones be formed and which item should be assigned to which zone? Answers to these questions can be given by solving corresponding optimization problems. Since both problems are strongly interrelated, we developed an integrative decision model that simultaneously determines zone size and item-to-zone assignment. Based on a full-factorial numerical study, we found that the model correctly reflects the planning situation. Due to its high complexity, the model can only be solved exactly for very small problem instances in an acceptable time. This motivates the use of solution approaches, which achieve near-optimal solutions with reasonable computational effort.
In the intended paper, we adapt the fix-and-optimize approach to solve the integrative zoning and item-to-zone assignment problem heuristically. Starting from an initial solution, the heuristic iteratively solves different sub-problems where a part of the decision variables is fixed. We analyze how the performance of the heuristic can be improved by generating the initial solution, specifying a fixation strategy, and limiting the number of iterations. To this end, we develop variants of the fix-and-optimize approach and conduct a numerical study in which the performance of the variants is compared with each other and with the exact approach.