In order to reduce waste collection costs and realize sustainable urban development, this paper investigates a two-echelon collaborative waste collection vehicle routing problem (2E-CWCVRP), considering the cooperation and profit distribution between participants in the collection network. An optimization model for 2E-CWCVRP with the aim to minimize total costs and carbon emissions is constructed. Then, a three-stage solution approach is developed to solve this model, including a k-means clustering for simplifying the problem, and a hybrid heuristic for searching the optimal vehicle routes based on Clarke & Wright algorithm and an adaptive large neighborhood search algorithm (CW-ALNS). Finally, an improved Shapley value model is constructed for determining the costs and carbon emissions reduction amount and the best alliance sequence of each participant. The experiment results indicate that: (1) the effectiveness of CW-ALNS algorithm is verified through the benchmark instances; (2) the costs and carbon emissions of the collection network could be reduced simultaneously after the implementation of cooperation; (3) constructing a large collection and transfer network is more efficient than dividing the network into several individual parts. Finally, different alliance sequences are analyzed from the economics and environment perspective and the best alliance sequences are determined.

为了降低废物收集成本并实现可持续的城市发展，本文考虑了收集网络参与者之间的合作和利益分配，研究了两级协作式废物收集车辆路由问题（2E-CWCVRP）。建立了用于2E-CWCVRP的优化模型，其目的是将总成本和碳排放量降至最低。然后，开发了一种三阶段解决方案方法来解决该模型，包括用于简化问题的k-均值聚类以及基于Clarke＆Wright算法和自适应大邻域搜索算法的用于搜索最佳车辆路线的混合启发式算法（ CW-ALNS）。最后，建立了改进的Shapley值模型，用于确定成本和碳减排量以及每个参与者的最佳联盟顺序。实验结果表明：（1）通过基准实例验证了CW-ALNS算法的有效性。（2）实施合作后，可以同时降低收集网络的成本和碳排放；（3）构建大型的收集和传输网络比将网络分成几个单独的部分更为有效。最后，从经济学和环境角度分析了不同的联盟序列，并确定了最佳联盟序列。（2）实施合作后，可以同时降低收集网络的成本和碳排放；（3）构建大型的收集和传输网络比将网络分成几个单独的部分更为有效。最后，从经济学和环境角度分析了不同的联盟序列，并确定了最佳联盟序列。（2）实施合作后，可以同时降低收集网络的成本和碳排放；（3）构建大型的收集和传输网络比将网络分成几个单独的部分更为有效。最后，从经济学和环境角度分析了不同的联盟序列，并确定了最佳联盟序列。