Premium
Solving the multidepot vehicle routing problem with limited depot capacity and stochastic demands
Author(s) -
Calvet Laura,
Wang Dandan,
Juan Angel,
Bové Lluc
Publication year - 2019
Publication title -
international transactions in operational research
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.032
H-Index - 52
eISSN - 1475-3995
pISSN - 0969-6016
DOI - 10.1111/itor.12560
Subject(s) - vehicle routing problem , depot , mathematical optimization , computer science , set (abstract data type) , metaheuristic , routing (electronic design automation) , operations research , monte carlo method , mathematics , computer network , statistics , archaeology , history , programming language
When capacity constraints have to be considered at each depot due to the existence of a limited number of capacitated vehicles, the multidepot vehicle routing problem (MDVRP) is a nontrivial extension of the well‐known capacitated vehicle routing problem (CVRP). The MDVRP combines a customer‐to‐depot assignment problem with several CVRPs. In real‐life scenarios, it is usual to find uncertainty in the customers' demands. There are few works on the stochastic MDVRP (SMDVRP) and, to the best of our knowledge, most of them assume the existence of an unlimited fleet of vehicles at each depot. This paper presents a simheuristic framework combining Monte Carlo simulation with a metaheuristic algorithm to deal with the SMDVRP with limited fleets (and, therefore, limited depot serving capacity). Its efficiency is tested on a set of stochastic instances, which extend the ones available in the literature for the deterministic version of the problem.