First Competitive Ant Colony Scheme for the CARP

Abstract

The CARP consists on determining a set of vehicle trips with minimum total cost. Each trip starts and ends at the depot, each required edge of the site graph is serviced by one single trip, and the total demand handled by any vehicle does not exceed the capacity Q. The most efficient metaheuristics published so far are a sophisticated taboo search method (CARPET) of Hertz et al. [1] and a hybrid genetic algorithm proposed by Lacomme et al. [2]. In the beginning, no collective memory is used and ants use only heuristic information. Pheromone deposition is proportional to the fitness that can be defined for minimization objectives as the inverse of the solution quality or solution cost. Local search can also be applied to increase the convergence rate. The process is iterated until a lower bound is reached or a maximal number of iterations is carried out.