Dynamic weather window-based multi-objective path planning for low-altitude logistics

Abstract This paper addresses the path planning problem of low-altitude logistics UAVs in dynamic weather environments by proposing a conceptual framework termed the "dynamic weather window". On this basis, a multi-objective optimization model incorporating delivery timeliness, operational energy consumption, and weather risk exposure is constructed, and a solution algorithm based on NSGA-III and rolling horizon re-planning is designed. Through simulation experiments, the proposed method improves the task completion rate from 19–78% achieved by baseline methods to 98% under severe weather scenarios, reduces the constraint violation rate from 22%-81% to 2%, and achieves a Hypervolume 38% higher than the optimal baseline. The research demonstrates that elevating weather from a "static constraint" to a "dynamic spatiotemporal window" can significantly enhance the adaptability of path planning to time-varying weather conditions; the rolling horizon re-planning mechanism strikes a balance between prediction and reaction, serving as an effective engineering approach to weather uncertainty; and the multi-objective Pareto front provides flexible safety-efficiency trade-off space for operational decision-making. The theoretical contributions of this paper lie in extending the theoretical boundary of the vehicle routing problem, establishing a coupled weather-flight performance relationship model, and providing a methodological paradigm that integrates multi-objective optimization with rolling horizon control.