A Vehicular Waiting Time Heuristic for Dynamic Vehicle Routing Problem

In the last years, the Vehicle Routing Problem (VRP) has been receiving increasing a special attention due its importance in the distribution of products. Considering demands that can appear during the service day, this work solves the Dynamic VRP (DVRP) with support of the environment simulator, capable to emulate the behavior of the vehicular fleet in a geographic 2D space. This work presents two main contributions: (i) the environment simulator that can be used to standardize DVRP test cases; and (ii) a Vehicular Waiting Time Heuristic (VWTH), that delays the attendance of demand with the meaning of reducing the total distance traveled by vehicular fleet. The proposed heuristic is tested by four optimization algorithms (Simulated Annealing, GRASP, Memetic Algorithm and Hill Climbing), and submitted to statistical tests for better conclusions. The tests showed, with 95% of confidence, that the VWTH is able to improve significantly the metaheuristics results.