Cargo theft weighted vehicle routing problem: modeling and application to the pharmaceutical distribution sector

Abstract

We present a variant of the weighted vehicle routing problem to design distribution routes in high theft risk areas. The model aims to minimize transportation and cargo theft costs. It uses a set of flow decision variables to calculate cargo value in each edge. Theft probability is calculated for each customer according to the historical frequency of theft occurrences reported in each location. The inclusion of theft costs derives in not so obvious route solutions as it might be less costly to drive longer distances or use more vehicles in order to minimize the value of cargo arriving to risky areas and therefore cargo theft losses. We adopt a simulated annealing metaheuristic as solution method. The model is tested in a real Brazilian pharmaceutical distribution company, highlighting the benefits of preventive routing in relation to theft occurrences.