A parallel multi-start NSGA II algorithm for multiobjective energy reduction vehicle routing problem

Abstract

The Multiobjective Energy Reduction Vehicle Routing Problem is a variant of the classic Vehicle Routing Problem where simultaneous optimization of more than one objective functions is required. In this paper, the problem is formulated with three different competitive objective functions. The first objective function corresponds to the optimization of the time needed for the vehicle to travel between two customers or between the customer and the depot, the second objective function is the minimization of the distance and the fuel consumption when a delivery route is planned and the third objective function is the minimization of the distance and the fuel consumption when a pickup route is planned. The problem is solved with a modified version of the NSGA II, with a use of more than one population, a multi start method for the creation of the initial population and a Variable Neighborhood Search algorithm for the improvement of the solution of each individual separately. In order to give the quality of the methodology, experiments are conducted using appropriately modified for the Vehicle Routing Problem instances based on the classic Euclidean Traveling Salesman Problem benchmark instances taken from the TSP library.