不确定条件下基于无人机的动态交通监控随机弧库存路由策略
Dynamic UAV-based traffic monitoring under uncertainty as a stochastic arc-inventory routing policy
摘要：
鉴于无人机的快速发展，以及对城市交通监控需求的日益增长，目前的研究空白之一是如何在多个时期系统地部署无人机。我们提出了一个实时数据驱动的方法：我们制定了第一个确定性弧库存路由问题，并推导出其随机动态策略。该政策预计将在不确定性最高和成本最高的情况下发挥最大价值，例如重大活动期间的城市交通监控。Bellman方程的近似所提出的库存路径策略制定为一个选择性的车辆路径问题。我们提出了一个近似的动态规划算法的基础上最小二乘蒙特卡罗模拟找到该政策。该算法已被修改，使最小二乘因变量被定义为“预期缺货成本下一次补货”。新算法进行了测试的30个模拟实例的实时轨迹超过5个时间段的选择性车辆路径规划，以评估所提出的政策和算法。对所选实例的计算结果表明，该算法可以优于近视政策的23%至28%，在这些测试，这取决于参数设计。进一步的测试进行经典基准弧布线问题的实例。11个链路的实例gdb19扩展成一个连续的15个周期的随机动态的例子，并用来证明为什么一个天真的静态多周期部署计划将是无效的，在真实的网络。
英文摘要：
Given the rapid advances in unmanned aerial vehicles, or drones, and increasing need to monitor traffic at a city level, one of the current research gaps is how to systematically deploy drones over multiple periods. We propose a real-time data-driven approach: we formulate the first deterministic arc-inventory routing problem and derive its stochastic dynamic policy. The policy is expected to be of greatest value in scenarios where uncertainty is highest and costliest, such as city traffic monitoring during major events. The Bellman equation for an approximation of the proposed inventory routing policy is formulated as a selective vehicle routing problem. We propose an approximate dynamic programming algorithm based on Least Squares Monte Carlo simulation to find that policy. The algorithm has been modified so that the least squares dependent variable is defined to be the "expected stock out cost upon the next replenishment". The new algorithm is tested on 30 simulated instances of real time trajectories over 5 time periods of the selective VRP to evaluate the proposed policy and algorithm. Computational results on the selected instances show that the algorithm can outperform the myopic policy by 23% to 28% over those tests, depending on the parametric design. Further tests are conducted on classic benchmark arc routing problem instances. The 11-link instance gdb19 is expanded into a sequential 15-period stochastic dynamic example and used to demonstrate why a naive static multi-period deployment plan would not be effective in real networks.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
