Algoritmo memético con operadores de inteligencia artificial para el CARP con inicio y fin no determinado y bi-objetivo

Abstract The arc routing problem with a variable starting/ending position (Open Capacitated Arc Routing Problem - OCARP), in its classic version, pursues the best strategy to serve a set of customers located in the network arcs using vehicles. Compared to the Capacitated Arc Routing Problem (CARP), the OCARP lacks of constrains that guarantee that each vehicle ought to start and end the tour at a given vertex (also known as a depot). The aim of this paper is to propose a heuristic to find an efficient frontier for the main objective functions: minimize the number of vehicles and the total cost. Additionally, a hybrid algorithm that complements the genetic algorithm with artificial intelligence operator is proposed.