Locating Automated Parcel Lockers (APL) with known customers’ demand: a mixed approach proposal

Abstract

Logistics Service Providers (LSP) are increasingly adopting Automated Parcel Lockers (APLs) to mitigate the operational pressure of last-mile logistics. The optimal location of APL stations is key for reaching customers’ demand while keeping the investment reasonable. Previous studies developed optimization algorithms and applied them to virtual instances of the problem, lacking applicability to real-life situations encountered by LSPs who aim to serve an urban area with such technology. This study proposes a novel solution to the APLs location problem by combining mixed-integer linear programming and greedy heuristics algorithms. The study tested the propose solution on real customers’ demand data related to Turin, Italy. Results show that covering 90% of the estimated potential demand requires 10 to 11 APLs, on average. The adopted approach enables finding an optimal solution grounded in a real geographical context without requiring time-consuming optimization.