Distribution network design in Large-Scale Retail Trade industry with K-means approach

The facility location problem, which aims to identify the best location to maximize or minimize a specific objective function, is among the most studied problem in the operations and logistics research literature. In the largescale retail trade (LSRT) industry, the distribution centres (DCs) location choice represents one of the most critical decision to design an efficient distribution network; indeed, this strategic decision has significant effects on logistic costs as well as on the effectiveness of distribution activities. In this paper we present a method for solving the facility location problem in the LSRT industry using a combined k-means clustering cascaded with a local search algorithm, finding the optimal balance between distribution costs on primary and secondary routes through 3 DCs. The approach has been validated on the real case of a LSRT company operating on the entire Italian territory.