Abstract: In this article, joint location-assignment problem is studied for designing an Urban Logistics Network (ULN) with multiple intermediate depots (IDs) and terminals of ULN. The key decisions are where to locate the intermediate depots (IDs) and how to assign terminals to IDs such that the total network cost is minimized. A large-scale static and deterministic integer programming model is presented solving a joint location-assignment problem of large chain retailers. To solve this model, a more efficient hybrid algorithm integrated with Genetic algorithm and Tabu search algorithm is put forward based on heuristic searching criteria. Our methodology is illustrated with the Urban Logistics Network from a leading Chinese retailer (Suning) in Chongqing. Numerical analysis suggests that optimal solutions can reduce the total network cost between 2.92%~14.36% by comparison with the current ULN.Delivery costs of "Last Mile" account for total urban logistics network costs more than 60%. Vehicle type is a deterministic factor to total urban logistics network costs. The effects of changing parameter values on the optimal solutions are also studied and some management implications are pointed out.

Key words: online-offline channel integration, iocation-assignment problem, “last mile”, Genetic algorithm, tabu search algorithm