New version of BHCVRP

Abstract

Vehicle Routing Problems involve a fleet of vehicles, seek to satisfy the demands of geographically dispersed customers and aim to minimize costs as long as existing constraints are met. New practical situations occur daily in the world, which leads to the increase of these problems and the emergence of new variants. Some of the solution methods are construction heuristics, which aim to quickly and easily find a first solution close to the optimum. However, there is the drawback that the vast majority of heuristics can only be used in the basic variant. In addition, there are currently few libraries that implement heuristic methods to solve these problems.

In this context, BHCVRP arises with the purpose of providing solutions to four types of vehicle route planning problems by employing seven construction heuristics. This paper presents a new version of this library implemented in Java and Python programming languages, which solves some of the deficiencies found in the previous version. Two new construction heuristics and three new variants are incorporated.Finally, from an experiment it is determined that the versions of BHCVRP do not show significant differences.

Keywords: BHCVRP, construction heuristics, software design, Vehicle Routing Problem.