天气不确定性下无人机配送的两阶段协同决策:提前取消与实时响应OACHSSCD

To address the challenges of unmanned aerial vehicle(UAV)delivery under weather disturbances,such as high susceptibility to interruptions,high cancellation costs,and reduced service reliability,this paper develops a decision-making optimization framework that supports both pre-cancellation in the planning stage and real-time cancellation during the execution stage.The framework aims to minimize total operating costs and enhance system resilience under uncertain weather environments.A two-stage stochastic programming model is established,incorporating key operational constraints such as UAV endurance,payload capacity,and multi-depot coordination.To solve large-scale stochastic vehicle routing problems,this paper designs a Heuristic Column Generation(HCG)algorithm,employs a Copula-based scenario generation method to capture the correlation between depot closures and customer inaccessibility caused by weather,and validates the effectiveness of the proposed model and algorithm through extensive numerical experiments.The results show that HCG maintains solution quality within 1%of exact benchmarks across all test instances while reducing computational time by two orders of magnitude,demonstrating strong scalability.Introducing proactive pre-cancellation significantly reduces real-time failure costs during the execution stage,and improves system resilience and service reliability,especially under high demand or capacity constraints.The interaction among capacity thresholds,endurance limits,and correlated weather disturbances explains the non-monotonic cost patterns observed in large-scale instances.Accordingly,operators should adopt a dual risk control mechanism of"proactive defense-dynamic stop-loss"and allocate resources based on weather correlation characteristics to achieve an optimal balance between cost efficiency and service reliability under uncertainty.