Routing for simultaneous pickup and delivery of goods

This project addressed the problem of routing and task-scheduling of robots for simultaneous pickup and delivery of goods, in simulation. I studied and compared the performance of metaheuristic algorithms on Vehicle Routing problem (VRP) variants. In particular, I applied Genetic and Greedy algorithms to the Capacitated VRP and VRP with time windows. I demonstrated a 30% reduction in the delivery time, and hence energy consumption, of 12 vehicles to pick and deliver goods at 50 locations. The experiments were coded in Python 3.0. Slides for the project can be found here. Below is the optimized route for 12 vehicles used to pickup and deliver items at 20 locations, with each color representing a vehicle used. Here, Greedy algorithm was used on the Capacitated VRP.