Fine-tuning a parametric Clarke and Wright heuristic by means of EAGH (empirically adjusted greedy heuristics)

Altınel and Öncan (2005) (A new enhancement of the Clarke and Wright savings heuristic for the capacitated vehicle routing problem) proposed a parametric Clarke and Wright heuristic to solve the capacitated vehicle routing problem (CVRP). The performance of this parametric heuristic is sensitive to fine-tuning. Antinel and Öncan used an enumerative parameter-setting approach and improved on the results obtained with the original Clarke and Wright heuristic, but their approach requires much more computation time to solve an instance. Battarra et al (2008) (Tuning a parametric Clarke–Wright heuristic through a genetic algorithm) proposed a genetic algorithm to set the parameter values. They succeeded in reducing the time needed to solve an instance, but the quality of the solution was slightly worse. In this paper, we propose to use the EAGH (empirically adjusted greedy heuristics) procedure to set the parameter values. A computational experiment shows the efficiency of EAGH; in an even shorter time, we improve on the best results obtained with any parametric Clarke and Wright heuristic method proposed in the literature.