Location-routing problem for customized product delivery considering assembly delay

Abstract

The delivery of assemble-to-order (ATO) products differs from that of standard products because the assembly process begins after receiving customer orders. As both assembly and distribution work have recently been outsourced to third-party logistics (3PL), the resulting preparation time gap between ATO and standard products imposes new time constraints on 3PL operations. To optimize delivery strategies and minimize total costs for 3PL providers under the new condition, this study addresses the location-routing problem for customized product delivery considering assembly delay (CPADLRP), which determines the location strategy of forward assembly warehouses and distribution strategy of ATO products simultaneously. We formulate the problem as a mixed-integer programming model, incorporating a novel variation of time window constraints with dynamically changing start time boundaries. An improved adaptive large neighborhood search (ALNS*) algorithm is proposed to solve this problem, incorporating a principal-component-analysis-based k-means and four problem-specific operators to the basic ALNS. Compared with the Gurobi Solver and two other typical heuristic algorithms, ALNS* achieves better solutions more efficiently. Moreover, we conduct a series of sensitivity analyses on the main factors of the problem and drive several managerial insights for 3PL providers to reduce total costs. The sensitivity analysis on the proportion of customized orders also verifies the importance of considering the assembly delay in the LRP.