Dealers GeoCluster
An R-based logistics tool that groups dealers geographically, assigns dealer shipments without exceeding truck capacity, and sequences each route by combining clustering with the Traveling Salesman Problem (TSP) and Capacitated Vehicle Routing Problem (CVRP).
The Optimization Problem
The network had two connected optimization problems: group nearby dealers into practical geographic clusters, then determine how trucks should serve them without exceeding vehicle capacity. That meant balancing dealer priority, shipment volume, route length, and the order of stops.
The Algorithm
Using R's geospatial libraries, I grouped dealers within a 100-mile working radius and weighted assignments by priority and unit demand. The routing layer then combined a Traveling Salesman Problem (TSP) sequence with Capacitated Vehicle Routing Problem (CVRP) constraints so each truck followed an efficient stop order while staying within load capacity. Interactive maps exposed cluster boundaries, dealer assignments, loads, and optimized routes for scenario testing.