Application of modified flower pollination algorithm to multiple vehicle routing problems with time constraints

Abstract

-The aim of this research is to apply the modified flower pollination algorithm (MoFPA) to the multiple vehicle routing problems (MVRP) with the time constraints. As one of the latest enhanced versions of the original flower pollination algorithm (FPA), the MoFPA utilizes the randomly switching probability for selection between local (self) and global (cross) pollinations in order to search for the better solutions and improve its search performance. In this paper, the MoFPA is applied to solve the MVRP problems with the time constraints based on the modern optimization approach. Ten standard MVRP problems with the time constraints consisting of approximately 100–500 destinations are selected and solved by the MoFPA. Results obtained by the MoFPA are compared with those obtained by original FPA, genetic algorithm (GA), tabu search (TS) and particle swarm optimization (PSO). Results reveals that the MoFPA can provide optimal solutions of all ten selected MVRP problems with shorter total distance than FPA, PSO, TS and GA, respectively. In addition, the execution time consumed by FPA and MoFPA are less than PSO, TS and GA, respectively.