随着垃圾分类政策的逐渐推广落实，且传统的垃圾回收模式在以大规模垃圾回收、垃圾分类储存运输、邻避效应、逆向物流碳排放、经济成本等为主的诸多方面已不能满足需求，我国对科学合理垃圾分类回收体系的需求逐渐迫切。本文为促进解决垃圾分类回收体系的建设问题，对实际工作提供参考、指导以及为相关领域进一步研究做出铺垫，从回收点的角度建立基于不同垃圾类别、邻避效应、逆向物流碳排放、费用等因素的选址-路径问题模型，设计相关方法、算法求解，并进行了实例应用。具体内容如下。

首先本文从实际问题出发进行数学建模，通过对比不同类型的回收设施，选择了具有较好前景的中大型回收设施进行规划，然后通过设定全局满意度、回收车辆遍历路径碳排放最小、回收点建设维护与路径运输费用三个目标函数，规定了解的形式与结构，完成对实际问题的抽象而得出多目标规划模型，其中最小碳排放目标为最短路目标的线性函数。考虑到模型的复杂性以及各目标之间权重比例无法确定而未采用传统多目标规划求解手段，而基于模型实际背景提出源头分类理念观点，阐述了该观点对于垃圾分类政策、理念发展的促进作用，随之提出与该理念相对应的模型分割求解方案，即将4种垃圾回收点分开讨论的同时对目标函数进行动态算法分阶段的求解方式，论证了该求解方案的可行性。

随后对应模型的分割求解方案，本文设计了两阶段的动态混合算法框架，外层对应满意度目标函数生成回收点选址初始解，内层得到选址初始解进行遍历路径的寻优，最后通过终止条件后，得到最优解及优质解集。基于良好全局搜索能力需要，外层算法选择了标准PSO，并基于Levy飞行随机原理对惯性权重w进行了动态化改进。通过动态化惯性权w重控制PSO能够满足在运行全过程中，处于不同时期有着不同全局寻优、局部寻优精确度侧重的需求。内层选择A*算法，基于其启发式原理完成了最短遍历路径寻优流程的实现。之后对内外层算法的输入输出接口、终止条件等进行了设计并完成动态混合算法的全流程。最后通过测试案例说明了算法的有效性。

在模型及其求解算法的应用层面，本文选择了西安市某区域。首先进行了需求点的数据采集与归一化，通过对需求点的K-means多维聚类、最小覆盖矩形包络等完成数据预处理，随后带入参数完成模型，并输入算法进行求解，整理得到基于回收点满意度与遍历最短回路的帕累托最优曲线。基于帕累托最优原理与目标函数最优化方向提取了多组备选优质解集，综合得到全局决策方案矩阵，引入BWM-TOPSIS方法对决策方案矩阵进行评价，得到基于设定决策情景下的最优全局决策方案，给出了该方案下的三个目标最优值与相关说明。最后依据实例的求解全流程给出了在不同决策情景下的求解方法与相关建议。

With the gradual promotion and implementation of waste classification policy, and the traditional waste recovery mode in many aspects such as large-scale waste recovery, waste classification storage and transportation, NIMBY effect, reverse logistics carbon emissions, economic costs and so on, has been unable to meet the demand, China's demand for scientific and reasonable waste classification and recovery system is gradually urgent. In order to promote the construction of waste classification and recycling system, provide reference and guidance for practical work, and pave the way for further research in related fields, this paper establishes a location routing problem model based on different waste categories, NIMBY effect, reverse logistics carbon emissions, cost and other factors from the perspective of recycling point, and designs relevant methods and algorithms to solve the problem, An example is given. The details are as follows.

 

First of all, this paper carries on the mathematical modeling from the actual problem, through comparing different types of recycling facilities, selects the medium and large-scale recycling facilities with good prospects for planning, and then sets three objective functions: global satisfaction, minimum carbon emission of recycling vehicle traversal path, recycling point construction and maintenance and Path Transportation cost, and specifies the form and structure of understanding, so as to complete the simulation A multi-objective programming model is obtained by abstracting the global problem, in which the minimum carbon emission goal is a linear function of the shortest path goal. Considering the complexity of the model and the uncertainty of the weight ratio between the objectives, the traditional multi-objective programming method is not used. Based on the actual background of the model, the concept of source classification is proposed, and the role of this concept in promoting the development of waste classification policy and concept is elaborated. Then, the solution scheme of model segmentation corresponding to this concept is proposed, that is, four kinds of waste collection points are separated At the same time, the objective function is solved by dynamic algorithm in stages, and the feasibility of the solution is demonstrated.

 

Then, corresponding to the segmentation solution of the model, this paper designs a two-stage dynamic hybrid algorithm framework. The outer layer generates the initial solution of the recycling point location according to the satisfaction objective function, and the inner layer obtains the initial solution of the location to optimize the traversal path. Finally, through the termination conditions, the optimal solution and high-quality solution set are obtained. Based on the need of good global search ability, the standard PSO is selected for the outer layer algorithm, and the inertia weight W is improved dynamically based on levy random flight principle. Through dynamic inertia weight W weight control, PSO can meet the requirements of different global optimization and local optimization accuracy in different periods in the whole process of operation. The inner layer selects A* algorithm and completes the shortest traversal path optimization process based on its heuristic principle. After that, the input and output interface and termination condition of the inner and outer algorithm are designed, and the whole process of the dynamic hybrid algorithm is completed. Finally, a test case is given to illustrate the effectiveness of the algorithm.

 

In the application level of the model and its solution algorithm, this paper selects an area in Xi'an. Firstly, the data collection and normalization of demand points are carried out, and the data preprocessing is completed by K-means multi-dimensional clustering and minimum coverage rectangular envelope of demand points. Then, the parameters are brought in to complete the model, and the input algorithm is used to solve the problem, and the Pareto optimal curve based on the satisfaction degree of recycling points and the shortest cycle is obtained. Based on Pareto optimization principle and objective function optimization direction, multiple sets of alternative high-quality solution sets are extracted, and the global decision scheme matrix is obtained. The BWM-TOPSIS method is introduced to evaluate the decision scheme matrix, and the optimal global decision scheme based on the set decision scenario is obtained. The three objective optimal values and related explanations under the scheme are given. Finally, according to the whole process of the solution of the example, the solution methods and relevant suggestions under different decision scenarios are given.

[1]杨帆, 邵超峰, 鞠美庭. 城市垃圾分类的国外经验[J]. 生态经济, 2016, 32(11): 2-5.
[2]金宜英, 邴君妍, 罗恩华,等. 基于分类趋势下的我国生活垃圾处理技术展望[J]. 环境工程, 2019, 37(09): 149-153+130.
[3]董飞, 扶漪红, 吴笑天, 等. 城市生活垃圾分类治理: 现实困境与实践进路[J]. 城市发展研究, 2021,28(02): 110-116.
[4]Dogan Ozgen, Bahadir Gulsun. Combining possibilistic linear programming and fuzzy AHP for solving the multi-objective capacitated multi-facility location problem[J]. Information Sciences, 2014, 268.
[5]Derya Dinler, Mustafa Kemal Tural, Cem Iyigun. Heuristics for a continuous multi-facility location problem with demand regions[J].
[6]张曦, 杨超, 黄松. 双目标多元需求的截流设施选址问题研究[J]. 公路交通科技, 2011, 28(02): 116-120+153.
[7]刘尚俊, 刘君. 一个具有随机模糊需求的多设施选址模型及解法[J]. 重庆工学院学报(自然科学版), 2009, 23(06): 71-73.
[8]陈中武. 银行网点选址优化问题研究[D]. 华中科技大学, 2013.
[9]姚冠新, 温伟锋. 应急物流多设施动态选址模型构建及检验[J]. 商业时代, 2013(10): 40-41.
[10]韦彤, 耿娜, 江志斌, 等. 面向确定需求的多服务多设施选址问题[J]. 工业工程与管理, 2014, 19(01): 47-52+84.
[11]陈义友, 陈以衡. 基于逐渐覆盖的自提点选址模型与算法研究[J]. 计算机应用研究, 2016, 33(08): 2275-2277+2297.
[12]胡大伟, 朱志强, 胡勇. 车辆路径问题的模拟退火算法[J]. 中国公路学报, 2006(04): 123-126.
[13]胡大伟, 陈诚. 遗传算法(GA)和禁忌搜索算法(TS)在配送中心选址和路线问题中的应用[J].系统工程理论与实践, 2007(09): 171-176.
[14]胡大伟, 陈诚, 郭晓汾. 带集货和配送的多站点VRP优化算法研究[J]. 数学的实践与认识, 2007(02): 98-104.
[15]胡大伟. 设施定位和车辆路线问题模型及其启发式算法研究[D]. 长安大学, 2008.
[16]胡大伟, 陈海妹, 梁一为, 等. 车辆与无人机混合编队的路径优化问题模型构建[J]. 长安大学学报(自然科学版), 2021, 41(01): 78-89.
[17]胡大伟, 刘成清, 胡卉, 等. 基于低碳视角的两阶段开放式选址路径问题——燃油车与电动车对比[J]. 系统工程理论与实践, 2020, 40(12): 3230-3242.
[18]Abdelhalim Hiassat, Ali Diabat,Iyad Rahwan. A genetic algorithm approach for location-inventory-routing problem with perishable products[J]. Journal of Manufacturing Systems, 2017, 42.
[19]Khosro Pichka, Amirsaman H. Bajgiran, Matthew E.H. Petering, et al. The two echelon open location routing problem: Mathematical model and hybrid heuristic[J]. Computers & Industrial Engineering, 2018, 121.
[20]邹云程. 基于充电选择的电动汽车充电设施布局以及充电收费优化研究[D]. 合肥工业大学, 2020.
[21]吕新福, 蔡临宁, 曲志伟. 废弃物回收物流中的选址-路径问题[J]. 系统工程理论与实践, 2005(05): 89-94.
[22]李想, 李苏剑, 李宏. 两级选址-路径问题的大规模邻域搜索模拟退火算法[J]. 工程科学学报, 2017, 39(06): 953-961.
[23]宋颂颂. 低碳化选址—路径问题优化模型研究[D]. 东北大学, 2012.
[24]张春苗, 赵燕伟, 张景玲, 等. 低碳定位——车辆路径问题[J]. 计算机集成制造系统, 2017, 23(12): 2768-2777.
[25]刘成清, 胡大伟, 黄榕. 基于路径灵活性的两阶段开放式低碳选址-路径问题[J]. 科学技术与工程, 2020, 20(17): 7080-7087.
[26]冷龙龙. 低碳物流选址路径问题及超启发算法研究[D]. 浙江工业大学, 2020.
[27]李贵华, 唐琦, 玄雪. 供应链管理环境下的逆向物流研究[J]. 物流科技, 2004(10): 63-65.
[28]龚翔, 朱万春. 低碳经济环境下物流业网络构建对城镇化发展的影响[J]. 生态经济, 2021, 37(03): 101-105.
[29]陈燕, 于放, 田月, 等. 基于Hadoop的车辆调度算法优化及应用[J]. 计算机系统应用, 2018, 27(10): 268-272.
[30]梁田, 曹德欣. 基于莱维飞行的改进简化粒子群算法[J/OL]. 计算机工程与应用: 1-14[2021-04-05]. http://kns.cnki.net/kcms/detail/11.2127.TP.20201030.0959.006.html.
[31]华勇, 王双园, 白国振, 等. 基于惯性权值非线性递减的改进粒子群算法[J]. 重庆工商大学学报(自然科学版), 2021, 38(02): 1-9.
[32]Ruba Abu Khurma, Ibrahim Aljarah, Ahmad Sharieh. A Simultaneous Moth Flame Optimizer Feature Selection Approach Based on Levy Flight and Selection Operators for Medical Diagnosis[J]. Arabian Journal for Science and Engineering, 2021(prepublish).
[33]谭威, 胡新荣, 雷伟. 基于Unity的A-Star算法在游戏中的具体实现[J]. 计算机产品与流通, 2018(11): 121.
[34]曾博, 徐富强, 刘一贤, 等. 综合考虑经济-环境-社会因素的多能耦合系统高维多目标规划[J/OL]. 电工技术学报: 1-12[2021-04-06]. https://doi.org/10.19595/j.cnki.1000-6753. tces.L90096.
[35]Woo Soomin, Bae Sangjae, Moura Scott J.. Pareto optimality in cost and service quality for an Electric Vehicle charging facility[J]. Applied Energy, 2021, 290.
[36]Lahri Vijay, Shaw Krishnendu, Ishizaka Alessio. Sustainable supply chain network design problem: Using the integrated BWM, TOPSIS, possibilistic programming, and ε-constrained methods[J]. Expert Systems With Applications,2021,168.
[37]Dong Jiuying, Wan Shuping, Chen Shyi Ming. Fuzzy best-worst method based on triangular fuzzy numbers for multi-criteria decision-making[J]. Information Sciences, 2021, 547.
[38]Jafar Rezaei, Thomas Nispeling, Joseph Sarkis, et al. A supplier selection life cycle approach integrating traditional and environmental criteria using the best worst method[J]. Journal of Cleaner Production, 2016, 135.
[39]Farrokhizadeh Elmira, Seyfi-Shishavan Seyed Amin, Kutlu Gündoğdu Fatma, et al. A spherical fuzzy methodology integrating maximizing deviation and TOPSIS methods[J]. Engineering Applications of Artificial Intelligence, 2021, 101.
[40]唐红林. “邻避效应”问题根源及对策分析[J]. 人民论坛, 2020(30): 64-65.
[41]蒋海青. 开放式低碳选址—路径模型及其算法研究[D]. 浙江工业大学, 2019
[42]张金良, 李超. 碳排放影响下的动态配送车辆路径优化研究[J/OL]. 中国管理科学: 1-13[2021-04-06]. https://doi.org/10.16381/j.cnki.issn1003-207x.2019.0816.
[43]温喜梅, 朱兴林, 刘泓君. 基于碳排放量对城市道路车辆路径规划研究[J]. 交通科技与经济, 2021, 23(02): 21-28.
[44]李文霞, 张春民, 马昌喜. 多目标低碳车辆路径优化模型及求解算法[J]. 交通信息与安全, 2020, 38(01): 118-126+144.
[45]刘丽姣. 考虑碳排放的多车场多车型VRP模型及算法研究[D]. 深圳大学, 2017.
[46]冯茜, 李擎, 全威, 等. 多目标粒子群优化算法研究综述[J/OL]. 工程科学学报: 1-10[2021-04-06]. https://doi.org/10.13374/j.issn2095-9389.2020.10.31.001.
[47]庞明礼, 于珂. “先行先试”的次优绩效及其扩散机制——以W市生活垃圾分类政策为例[J]. 地方治理研究, 2020(01): 2-12+78.
[48]Payam Shojaei, Seyed Amin Seyed Haeri, Sahar Mohammadi. Airports evaluation and ranking model using Taguchi loss function, best-worst method and VIKOR technique[J]. Journal of Air Transport Management, 2018, 68.