Metamodel-Based Optimization of the Article-to-Device Assignment and Manpower Allocation Problem in Order Picking Warehouses

Abstract

Efficient order picking requires a coordinated way of combining and utilizing three kinds of heterogeneous resources: articles, devices, and operators. Usually, the assortment of articles is subject to permanent adaptations. Hence, the interdependent decisions of assigning articles to devices and allocating manpower among devices need to be adjusted and the problem has to be solved frequently for similar instances. We propose a combination of exact and heuristic solution approaches. For an immediate reaction to each assortment change, a heuristic approach applying metamodel-based optimization is used. The data required for estimating the metamodel is provided by an exact approach which is utilized from time to time to reset the system to an optimal state. Based on sampled data of a pharmaceutical wholesaler, we compare exact and heuristic approach with regard to quality and time of solving in-sample and out-of-sample instances.