Abstract: Considering the current community group purchase fresh products cold chain transportation process, due to quality decay leading to customer satisfaction is reduced, while the demand blowout causes insufficient capacity, a two-level cold chain collaborative optimization distribution strategy based on the crowdsourcing model is proposed, that is, the enterprise refrigerated truck completes the first-level cold chain transportation from the city warehouse to transit warehouse, and the crowdsourced refrigerated truck completes the secondary cold chain transportation from the transfer warehouse to the head of the regiment, and the total cost including service delay cost, carbon emission cost and fixed cost is minimized. And the customer’s satisfaction with the product quality is the optimization goal, and a two-stage open-close hybrid cold chain vehicle low-carbon path planning model with crowdsourcing is established. According to the characteristics of the model, an improved adaptive large neighborhood search (IALNS) algorithm is constructed, a new destruction-repair solution strategy is designed, and the idea of simulated annealing (SA) is added in the operator selection stage to speed up the convergence speed and improve the global search capability of the algorithm. By comparing the results with the example optimization results of adaptive large neighborhood search (ALNS) , simulated annealing (SA) , genetic algorithm (GA) , and particle swarm optimization (PSO) , it is proved that this method is feasible and effective. The strategy takes into account corporate profits and customer needs, and compares the experimental results in different distribution modes, which verifies that the model has positive significance in solving the problem of cold chain logistics of community group purchase fresh products.

Key words: two-echelon cold chain planning, time window, community group purchase, low carbon, crowdsourcing, improved adaptive large neighborhood search

摘要： 考虑当前社区团购生鲜品冷链运输过程中，因品质衰减导致顾客满意度降低，同时需求井喷造成运力不足的问题，提出一种基于众包模式的两级冷链协同优化配送策略，即由企业冷藏车完成城市仓至中转仓的一级冷链运输，由众包冷藏车完成由中转仓至团长的二级冷链运输，并以包含服务延迟成本、碳排放成本和固定成本的总成本最小，及团长对产品品质满意度最大为优化目标，建立一个带有众包的两级开闭混合冷链低碳路径规划模型。针对模型特点，构造了一种改进的自适应大领域搜索算法（improved adaptive large neighborhood search，IALNS），设计了新的破坏-修复解的策略，并在算子选择阶段加入模拟退火算法（simulated annealing，SA）的思想，以加快收敛速度，提高算法全局搜索能力。通过分别与自适应大领域算法（adaptive large neighborhood search，ALNS）、模拟退火算法、遗传算法（genetic algorithm，GA）、粒子群优化算法（particle swarm optimization，PSO）的算例优化结果对比，证明该算法的有效性。该策略兼顾企业利润和客户需求，对比在不同配送模式下的实验结果，验证了该模型在解决社区团购生鲜品冷链物流问题上有积极意义。

关键词: 两级冷链规划, 时间窗, 社区团购, 低碳, 众包, 改进的自适应大领域搜索算法