A novel bi-objective credibility-based fuzzy model for municipal waste collection with hard time windows

Abstract

In this study, a bi-objective vehicle routing mathematical model was proposed, in which the main objectives were minimizing the total economic cost, as well as the whole time for municipal waste collection. The considered waste collection network included bins, multiple depots, multiple heterogeneous vehicles, multiple intermediate facilities, and a landfill. Besides, in each day, municipal waste was collected several times in various hard time windows by heterogeneous vehicles, which all used vehicles could trip more than once per time window. It should be noted that the amount of waste generated was uncertain, which a fuzzy credibility theory was used to cope with this uncertainty. The exact solutions of some small problems were generated by the augmented ε-constraint method. Due to the complexity of this problem in more substantial sizes, the metaheuristic of Non-dominated Sorting Genetic Algorithm II (NSGA-II) was used. Since the performance of metaheuristic algorithms is quite sensitive to their parameters, we used the self-adaptive method to tune the parameters of NSGA-II and also compared the results with the Taguchi method. The initial solutions of the metaheuristic approach were generated by a novel heuristic algorithm. Also, the waste collection vehicle routing problem of one of Tehran’s regions was solved by the proposed model. After solving this model, the presented Pareto optimal solutions showed the total economic cost and the total time of the waste collection were improved by about 1.4% and 1.1%, respectively, which means a significant reduction, due to the high volume of costs and time.