...
  • 主题
高级搜索  >
历史搜索 清空历史
首页> 外文期刊>Quality Control and Applied Statistics >Improving quality of prediction in highly dynamic environments using approximate dynamic programming
【24h】

Improving quality of prediction in highly dynamic environments using approximate dynamic programming

机译：使用近似动态编程提高高度动态环境中的预测质量

获取原文
获取原文并翻译 | 示例
           
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

Purpose:To present an approximate dynamic programming approach to taxi-out time of a flight prediction which is an indicator of surface congestion.Summary:In air traffic control, the departure clearance decisions depend on Current situational awareness, which is: the number of aircrafts that are taxiing toward the runways, airport weather, pending clearance requests, and runway configurations, and Also predictions made on impending congestions caused by high variance in the demand for arrivals and departures that are triggered by en route delays.While current situational awareness can be obtained via real-time monitoring, the challenge lies in accurately predicting the surface congestions which can help in optimized departure decision making. Learning-based approximate dynamic programming (ADP) methods are suited for sequential decision making to make predictions in highly dynamic environments. The prediction problem is cast in a probabilistic framework of stochastic dynamic programming and solved using ADP approaches especially reinforcement learning (RL). The paper applies the RL methodology to test its effectiveness in a highly dynamic taxi-out time prediction problem in which taxi-out time is predicted 15 minutes before gate-departure time of a flight. (41 refs.)
机译：目的：为飞行预测的滑行时间提供一种近似的动态编程方法，这是表面拥挤的指标。摘要：在空中交通管制中，离场许可决定取决于当前的态势感知，即：飞机数量滑行向跑道，机场天气，待批准的通关请求和跑道配置，以及对因途中延误引发的进出港需求差异很大而导致的即将发生的拥堵做出的预测。通过实时监控获得的挑战在于准确预测表面拥堵，这将有助于优化出发决策。基于学习的近似动态编程（ADP）方法适用于顺序决策，以便在高度动态的环境中进行预测。预测问题在随机动态规划的概率框架中提出，并使用ADP方法（尤其是强化学习（RL））解决。本文采用RL方法来测试其在高度动态的滑行时间预测问题中的有效性，该问题中，滑行时间是在航班登机口起飞时间前15分钟进行预测的。 （41篇）

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号