Abstract: In order to optimize the distribution route of hazardous materials in the city, firstly, considering the uncertain distribution of traffic flow near the route and the impact of traffic volume changes on the transportation risk, a dynamic uncertainty risk assessment model is designed. Secondly, according to the low-carbon requirements of green distribution, the carbon emissions under the conditions of vehicle load and speed changes are calculated. A multi-objective nonlinear optimization model is established with the goal of minimizing transportation risk, transportation cost and carbon emission, and considering the time window and the load limit. Based on model feature analysis, an improved NSGA-II is designed. In the initial stage of the population, random perturbation operation is added, an elite selection strategy based on threshold update is set in the selection process, and an individual negative selection strategy and a variable neighborhood search strategy are designed to improve the algorithm optimization ability. Finally, combined with practical examples to solve, the results show that the improved algorithm has good solution performance. The research results can provide decision support for government departments and transportation enterprises to realize the safe, economical and green distribution of hazardous materials in cities.

Key words: hazardous materials transportation, distribution route optimization, dynamic uncertainty risk, green distribution, NSGA-II, variable neighborhood search

摘要： 为实现城市内的危险品配送路径优化，考虑路线附近交通流量的不确定分布及运量变化对运输风险的影响，设计了动态不确定风险评估模型。根据绿色配送的低碳要求，对车辆负荷及速度变化等条件下的碳排放量进行计算。以最小化运输风险、运输费用以及碳排放为目标，并考虑时间窗与载重量限制，建立多目标非线性优化模型。基于模型特征分析，设计了改进的NSGA-II算法。在种群初始化阶段加入随机扰动操作，选择过程中设置基于阈值更新的精英选择策略，设计个体阴性选择策略与变邻域搜索策略以提高算法寻优能力。结合实际算例进行求解，结果表明：改进后的算法具有良好的求解性能。研究结果可为政府部门和运输企业实现城市内危险品的安全、经济、绿色配送提供决策支持。

关键词: 危险品运输, 配送路线优化, 动态不确定风险, 绿色配送, NSGA-II, 变邻域搜索