A multiple colonies artificial bee colony algorithm for a capacitated vehicle routing problem and re-routing strategies under time-dependent traffic congestion. (July 2017)

• Record Type: Journal Article
• Authors: Ng, K.K.H.; Lee, C.K.M.; Zhang, S.Z.; Wu, Kan; Ho, William
• Is Part Of: Computers & industrial engineering. Volume 109(2017)
• Page Start: 151
• Page End: 168
• Publication Date: 2017-07
• Subjects: Online vehicle routing problem -- Swarm intelligence -- Artificial bee colony algorithm -- Multiple colony strategy; Engineering -- Data processing -- Periodicals; Industrial engineering -- Periodicals; 620.00285
• DOI: 10.1016/j.cie.2017.05.004
• Languages: English
• ISSNs: 0360-8352

Abstract

Highlights: Two novel Artificial Bee Colony (ABC) algorithms for VRP are proposed. Real time traffic scenario has been considered in implementing dynamic VRP. Rerouting scheme is an option in VRP to enhance the routing flexibility. The robustness of the proposed algorithms outperforms other ABC algorithm. Multiple colonies strategy is adopted to solve premature convergence. Abstract: An Online Vehicle Routing Problem is a formation of Capacitated Vehicle Routing Problem with re-routing strategy to resolve the problem of inefficient vehicle routing caused by traffic congestion. A flexible delivery rerouting strategy is proposed, which aims at reducing the risk of late delivery. The method of terminating an exploration in a solution by the original ABC algorithm, when the solution is trapped in local optima, is to abandon the solution after specific tolerance limits are set. The phenomenon of local optimal traps will be repeated rapidly after a lengthy recursive process and will eventually result in a low quality solution, with a more complex combinatorial problem when the capability of the exploration is restricted by an inflexible termination criterion. Therefore, this paper proposes a novel scheme using a Multiple Colonies Artificial Bee Colony algorithm. The designs of the outstanding bee selection for colony communication show it to be superior in exploitation. The performance of the proposed algorithm is examined through by Capacitated Vehicle Routing instances and a case study, and the results indicate the potential of using real time information for data-driven vehicle scheduling.