基于无人机的高层住宅最后“一百米”配送优化

针对城市高层住宅顾客对上门配送服务的需求，借助无人机停放平台，考虑包裹异质性以及无人机在不同配送阶段的精确能耗，构建以无人机飞行成本和能耗成本最小为目标，以无人机容量、电池组容量等为约束的高层住宅无人机上门配送模型，解决“垂直位置最后一百米配送”问题。基于此模型，设计带变邻域下降(VND)搜索的混合蚁群算法(HACO-VND)，引入4个算子进行变邻域下降搜索，为了提高算法的求解性能，提出两种局部搜索算子组合，根据顾客点数量使用不同的算子组合。实验结果表明，HACO-VND算法较CPLEX在求解精度与求解时间方面更优，特别是在大中型算例中表现出较佳性能。参数分析表明，高层住宅楼层数越多，无人机单次 飞行的能耗利用率越大，无人机容量与电池组容量共同对配送方案产生影响。为以后无人机送货上门服务方面的研究提供参考和思路。

Optimization of Final "100 Meters" Drone Delivery in High-rise Residential Buildings

This paper focuses on optimization of door-to-door distribution path for high-rise residence. A high-rise residential drone door-to-door distribution model is constructed to minimize the flight cost and energy consumption cost of drone, to solve the problem of "distribution in the last 100 meters of vertical position". Considering the heterogeneity of packages and the energy consumption of drone, the constraints of the model involve drone capacity, battery pack capacity, and other factors. Based on this model, a hybrid ant colony optimization with VND (HACOVND) is designed, and four operators are introduced for variable neighborhood descent (VND) search. To improve the solution performance of the algorithm, two local search operator combinations are proposed, and different operator combinations are used for different number of customer points. The experimental results show that the HACO-VND algorithm performs better than CPLEX in terms of calculation accuracy and time, especially for large and mediumsized examples. Parameter analysis shows that the higher of the high- rise residential buildings, the greater the energy consumption utilization rate in a single flight for the drone. The drone capacity and battery pack capacity jointly affect the distribution scheme. The result of the study provides reference and ideas for the drone door-to-door service research in the future.