A Multi-objective Task-Driven Vehicle Routing Problem with Recirculating Delivery and its Solution Approaches

The task-driven vehicle routing problem with recirculating delivery (TDVRPRD) is similar to the expansion of the multi-objective multi-depot vehicle routing problem with time windows. Applied to the characteristics of large quantities of orders and high frequency round trip in real express delivery, the important problem is how to match task orders with delivery vehicles and make reasonable path planning for delivery vehicles. In this paper, we establish a task-driven vehicle recirculating distribution model with multi objectives of minimizing the total cost and maximizing the total profit of distribution, propose the heuristic algorithm and variable neighborhood search algorithm respectively for the existing traditional single delivery and recirculating delivery. The four algorithms are compared and tested for the simulation experimental data. We present a variety of results showing that the task-driven vehicle recirculating distribution model has strong superiority under weak time constraints, and the specific variable neighborhood search algorithm has high search quality and stability.