A Variable Neighborhood Search for the Capacitated Arc Routing Problem with Intermediate Facilities

In this talk we present a unified Variable Neighborhood Search (VNS) algorithm for the Capacitated Arc Routing Problem (CARP) and some of its variants the Capacitated Arc Routing Problem with Intermediate Facilities (CARPIF) and the Capacitated Arc Routing Problem with Intermediate Facilities and Length restrictions (CLARPIF). Routing problems represent one of the most prominent classes of models within computational logistics, and within combinatorial optimization in general. Being both of practical and theoretical relevance, they have provided a primary arena for validating new metaheuristics and for determining comparative efficiency. While the first application problem for a metaheuristics has usually been a node routing problem such as the travelling salesman problem (TSP) or the vehicle routing problem (VRP), several of these methods have also been applied to arc routing problems. The capacitated arc routing problem (CARP) originally proposed by Golden and Wong [3] is one of the prototypical routing problems, asking a fleet of vehicles to service a set of clients which are distributed on the arcs of a road network. While every CARP instance can be transformed into an equivalent VRP instance, this comes at the cost of obtaining a graph which has a number of nodes equal to the number of arcs of the original CARP graph, which is usually very large. Therefore, there is ample room for research that works directly on the CARP formulation.