A bi-level multi-objective location-routing model for municipal waste management with obnoxious effects. (November 2021)

• Record Type: Journal Article
• Authors: Ma, Yanfang; Zhang, Wen; Feng, Cuiying; Lev, Benjamin; Li, Zongmin
• Is Part Of: Waste management. Volume 135(2021)
• Page Start: 109
• Page End: 121
• Publication Date: 2021-11
• Subjects: Municipal waste management -- Location-routing problem -- Obnoxious effects -- Multiple objective bi-level programming -- Hybrid NSGA-II; BP bi-level programming -- BKS best-known solution -- CLRP capacitated location routing problem -- CW clarke and wright -- FLP facility location problem -- HGA hybrid genetic algorithm -- MWM municipal waste management -- NP non-deterministic polynomial -- NSGA-Ⅱ non-dominated sorting genetic algorithm-Ⅱ -- VRP vehicle routing problem; Hazardous wastes -- Periodicals; Refuse and refuse disposal -- Periodicals; 363.728
• DOI: 10.1016/j.wasman.2021.08.034
• Languages: English
• ISSNs: 0956-053X

Abstract

Highlights: A bi-level multi-objective location-routing model is proposed for MWM. The bilevel relationship between the government and the sanitation company is discussed. The obnoxious effect function of a waste recycling center is defined. Two initialization methods are adopted: Clustering for leader and CW for follower. Two classic LRP benchmark cases and a case in Tianjin are used to test the NSGA-II. Abstract: Municipal waste management is a complex problem. This paper develops a bi-level multi-objective location-routing model for municipal waste management that considers the interests of both the government and the sanitation companies. The government as the leader decides on the location and scale of the waste recycling centers to reduce the obnoxious effects and ensure cost effectiveness, and the sanitation company as the follower decides on the waste collection routing plans based on the government-approved locations to minimize the logistics cost. An improved hybrid NSGA-II is then developed to solve the proposed model. Two initial solution methods are employed: clustering for the leader and a Clarke and Wright method for the follower. Non-dominated sorting and best-cost route crossover operator are used to improve the effectiveness of NSGA-II. Based on Prins (24 instances) and Barreto (13 instances) benchmarks, the experimental results indicated that the improved operator had strong competitiveness and a better performance than previous methods, with the improved algorithm achieving the best average gaps of 0.18% and 0.24% and improving the best-known solutions in some instances. The model and solution methodology are illustrated using a waste collection problem in Tianjin, from which practical insights are derived.