Perishable food distribution in urban area based on real-road network graph

Since their introduction, Vehicle Routing Problems (VRPs) are increasingly receiving attention of researchers. Several variants of VRPs have been introduced for numerous applications. Most of the approaches found within the literature propose to address VRPs on a customer-based graph, considering that every two nodes are linked by a single best path computed based on one criterion that is mostly the distance. However, in reality, many features can be associated to each road segment, hence alternative paths should be considered. In this study we propose a mathematical model to deal with Vehicle Routing Problem under capacity and Time Window constraint with alternative Paths denoted by CVRPTW-P for perishable food delivery. To handle the issue aforementioned, we model the problem using real road network graph. In the proposed approach alternative paths are considered in order to design routes based on real spatial data instead of mathematical equations which is commonly used in the literature. Computational experiment was conducted using CPLEX solver for random instances. Results prove that exact resolutions fail to solve large scale instances due to model complexity.