THE NEW METHODOLOGY FOR VEHICULAR NETWORK WITH FUZZY TIME WINDOWS

Abstract

This work proposes the new methodology for the vehicular network with fuzzy time windows. The Fuzzy technique is applied to produce an initial population and then the evolutionary algorithm is employed to improve the solution. In this work, the inter-route crossover, intra-route mutation, elitism strategy, and onlooker bee probability selection method were enhanced in the original processes of the evolutionary algorithm. The proposed algorithm is tested on 56 datasets of Solomon. The results from the proposed algorithm are shown in comparison with other algorithms in the literature. The findings from the computational results are very inspiring, it shows that the algorithm is very competitive. Comparing with the algorithm in the literature, the proposed algorithm obtains the best solution in terms of the coefficient of variation values for almost 40 instances from the 56 problem instances. In addition, the information regarding the p-value was resolved by the Wilcoxon signed-rank test for the considered testing instances that display statistically superior performance at the 95% significance level (α = 0.05) on comparing algorithms.