基于鲁棒优化与多阶段组合拍卖的动态垃圾清运联动机制及算法研究
批准号:
72071093
项目类别:
面上项目
资助金额:
49.0 万元
负责人:
徐素秀
依托单位:
学科分类:
运筹与管理
结题年份:
2024
批准年份:
2020
项目状态:
已结题
项目参与者:
徐素秀
中文摘要
考虑城市商业垃圾清运分配及定价问题,对物流需求的动态性及联动性等核心属性进行凝练,将智能商业垃圾清运系统抽象为“动态垃圾清运联动(DWCS)”系统。研究面向DWCS的鲁棒优化方法与多阶段组合拍卖机制:“通过鲁棒优化方法,事前调整车辆调度等方案,前瞻性地缓解需求变动对DWCS机制的干扰;结合多阶段组合拍卖机制,事后调整资源配置等方案,后顾性地削减需求变动对系统绩效的影响”。第一,针对事前需求扰动,将鲁棒优化中的不确定集抽象为几类线性规划约束,并将其纳入车辆路径优化模型和资源配置模型;第二,针对需求联动性及事后需求扰动,构建基于VCG思想和前景理论的多阶段组合拍卖机制,保证机制的有效性和公平性;第三,针对资源配置-车辆路径双层NP-难问题,构建基于变邻域搜索(VNS)和禁忌搜索(TS)的混合启发式算法,保证算法的效率和质量。最后,项目成果拟在具有典型需求特征的大型商业垃圾清运企业进行应用测试。
英文摘要
This project considers the allocation and pricing problems of city commercial waste collection, in which the key features of logistics demand such as dynamics and synchronization are incorporated, and the intelligent commercial waste collection system is abstracted as the “dynamic waste collection synchronization (DWCS)” system. The robust optimization method and multi-stage combinatorial auction mechanism for DWCS are proposed; that is, “the impact of ex ante demand disturbance on the DWCS mechanism can be alleviated through a prior vehicle routing adjustment policy based on the robust optimization method, and meanwhile a real-time reallocation scheme based on the multi-stage combinatorial auction mechanism can reduce the influence of ex post demand disturbance on the overall system performance.” Firstly, regarding the ex ante demand disturbance problem, the uncertainty set in robust optimization is abstracted into several types of linear programming constraints and then incorporated into the vehicle routing model and resource allocation model. Secondly, regarding the ex post demand disturbance and synchronization problem, a multi-stage combinatorial auction based on VCG mechanism and prospect theory is designed to ensure the effectiveness and fairness. Thirdly, regarding the two-layer NP-hard allocation-routing problem, a hybrid heuristic algorithm based on variable neighborhood search (VNS) and tabu-search (TS) is constructed to ensure the efficiency and quality. Finally, the analytical and experimental results of this project will be extensively tested based on the data provided by the large commercial waste collection enterprises with typical demand characteristics.
本研究聚焦于城市商业垃圾清运的分配与定价问题,提出“动态垃圾清运联动(Dynamic Waste Collection Synchronization, DWCS)”系统,以应对物流需求的动态性与联动性。针对垃圾运输车队规模及路径优化问题,构建多阶段优化模型,采用多种方法优化运输车辆停车中心的选址,并保证碳中和的基础上,通过智能算法提升整体运输效率。针对垃圾运输服务资源分配与定价的双边和单边市场,设计快速交易的按需拍卖机制,解决运输时间不确定性问题,保证模型及机制的鲁棒性。针对垃圾仓储管理问题,优化资源分配并确保拍卖机制的可行性与鲁棒性。在电动汽车电池回收领域,结合ChatGPT与物联网技术,通过多阶段Vickrey-Clarke-Groves (VCG) 拍卖机制优化废旧电池的信息匹配与资源分配。为提升电子废弃物回收效率,结合企业社会责任与乡村振兴目标,设计多阶段VCG拍卖机制,研究政府补贴政策的影响。针对垃圾分类的样本挑战,基于鲁棒优化理论提出智能高效算法,增强模型泛化能力。研究成果已在相关企事业单位进行测试与验证。研究成果响应社会、环境和经济发展的需求,为构建以智慧城市为驱动的智能垃圾清运系统提供了坚实基础,展现出广阔的应用前景和实际价值。
国内基金
海外基金