Routing Optimization in Transportation Network Design

Delivery companies are facing increasing financial pressures and strict environmental regulations to enhance their transportation networks. The rise of e-commerce has heightened the need for timely and efficient parcel delivery services. This research focused on optimizing vehicle routing for mid-mile delivery, a key challenge in logistics. Existing literature overlooks the complexities associated with the unique characteristics of regional postal networks, including constraints related to time windows, vehicle capacity, and the interplay between first-mile, mid-mile, and last-mile logistics. The main research question defined was: “How to redesign postal services’ routing between DCFs and DOs to reduce transportation costs while meeting service standards?” A practical solution was developed to improve mid-mile delivery efficiency, reduce transportation costs, and lower emissions. Mixed-integer linear programming (MILP), combined with a heuristic algorithm, was employed to address routing and vehicle selection challenges. Key constraints of the model included maximum travel time, cut-off times for deliveries and collections, time windows for facility arrivals, and vehicle capacities. The study considered a mixed fleet with different capacities and costs, vehicle access limits at some facilities, and separate routes for delivery and pickup. The MILP formulation was implemented using CPLEX and benchmarked against a heuristic tool provided by Google. Empirical validation on a regional subnetwork of Canada Post offices in British Columbia showed significant improvement in truck assignments, utilizing distance and travel time data from Google’s Distance Matrix API. This research contributes a decision-support framework that enhances mid-mile delivery efficiency, reduces transportation costs, and helps postal organizations improve operations, customer satisfaction, and sustainability.