Mathematical modelling and heuristic approaches to the location-routing problem of a cost-effective integrated solid waste management

作     者：Asefi, H. Lim, S. Maghrebi, M. Shahparvari, S. 

作者机构：Univ New South Wales Sch Civil & Environm Engn Sydney NSW 2052 Australia Ferdowsi Univ Mashhad Dept Civil Engn Mashhad Iran RMIT Univ Sch Business IT & Logist Melbourne Vic 3000 Australia 

出 版 物：《ANNALS OF OPERATIONS RESEARCH》 (运筹学纪事)

年 卷 期：2019年第273卷第1-2期

页      面：75-110页

核心收录：

学科分类：1201[管理学-管理科学与工程(可授管理学、工学学位)] 07[理学] 070104[理学-应用数学] 0701[理学-数学] 

基　　金：University of New South Wales, UNSW Curtin University of Technology Shahid Beheshti University of Medical Sciences, SBUMS 

主　　题：Municipal solid waste Integrated solid waste management Location-routing problem Mixed-integer linear programming Simulated annealing Variable neighbourhood search 

摘      要：Integrated solid waste management (ISWM) comprises activities and processes to collect, transport, treat, recycle and dispose municipal solid wastes. This paper addresses the ISWM location-routing problem in which different types of municipal solid wastes are factored concurrently into an integrated system with all interrelated facilities. To support a cost-effective ISWM system, the number of locations of the system s components (i.e. transfer stations;recycling, treatment and disposal centres) and truck routing within the system s components need to be optimized. A mixed-integer linear programming (MILP) model is presented to minimise the total cost of the ISWM system including transportation costs and facility establishment costs. To tackle the non-deterministic polynomial-time hardness of the problem, a stepwise heuristic method is proposed within the frames of two meta-heuristic approaches: (i) variable neighbourhood search (VNS) and (ii) a hybrid VNS and simulated annealing algorithm (VNS+SA). A real-life case study from an existing ISWM system in Tehran, Iran is utilized to apply the proposed model and algorithms. Then the presented MILP model is implemented in CPLEX environment to evaluate the effectiveness of the proposed algorithms for multiple test problems in different scales. The results show that, while both proposed algorithms can effectively solve the problem within practical computing time, the proposed hybrid method efficiently has produced near-optimal solutions with gaps of 4%, compared to the exact results. In comparison with the current cost of the existing ISWM system in the study area, the presented MILP model and proposed heuristic methods effectively reduce the total costs by 20-22%.