IOE591_Project

Vehicle Routing Problem applied to grocery delivery at Ann Arbor

For this purpose we use the city of Ann Arbor with Kroger stores as the depot for our vehicles, and based on the residential areas we distribute the customers throughout the region. The goal is to determine the optimal route from the depots to each customer location which generates the minimum travel cost. Two approaches have been discussed here, one with vehicles being allocated through bin-packing applied to different items, and the other by assuming all item demands into a single overall demand. The problem is then divided into three parts - Single-depot VRP, Multidepot VRP and heuristic approach through k-Means clustering - to understand trade-offs between alternatives. Through our results we try to highlight what factors come into play in achieving a good objective value in each alternative and the choice of best performing model based on CPU runtime and achieved objective value.