arXivDaily arXiv每日学术速递 周一至周五更新
北美退伍军人节(Veterans Day),12 日无更新
arXiv 2108.00710cs.ROcs.AI
收藏

基于多目标路径的D* Lite

Multi-Objective Path-Based D* Lite

Carnegie Mellon University(卡内基梅隆大学);Texas A&M University(得克萨斯农工大学)

Zhongqiang Ren, Sivakumar Rathinam, Maxim Likhachev, Howie Choset

摘要:

增量图搜索算法(如D* Lite)重用以前的搜索(可能是部分搜索),以加快后续的路径规划任务。在这篇文章中,我们感兴趣的是开发增量图搜索算法的路径查找问题,同时优化多个目标,如旅行风险,到达时间等,这是具有挑战性的,因为在多目标设置,“帕累托最优”的解决方案的数量可以指数增长相对于图的大小。本文提出了一种新的多目标增量搜索算法,称为多目标路径为基础的D* Lite(MOPBD*),利用基于路径的扩展策略修剪支配的解决方案。此外,我们还引入了MOPBD* 的次优变体,以提高搜索效率,同时近似帕累托最优前沿。我们数值评估的性能MOPBD* 和它的变种在各种地图有两个和三个目标。结果表明,我们的方法是更有效的,从零开始搜索,并运行到一个数量级的速度比现有的增量多目标路径规划方法。

英文摘要:

Incremental graph search algorithms such as D* Lite reuse previous, and perhaps partial, searches to expedite subsequent path planning tasks. In this article, we are interested in developing incremental graph search algorithms for path finding problems to simultaneously optimize multiple objectives such as travel risk, arrival time, etc. This is challenging because in a multi-objective setting, the number of "Pareto-optimal" solutions can grow exponentially with respect to the size of the graph. This article presents a new multi-objective incremental search algorithm called Multi-Objective Path-Based D* Lite (MOPBD*) which leverages a path-based expansion strategy to prune dominated solutions. Additionally, we introduce a sub-optimal variant of MOPBD* to improve search efficiency while approximating the Pareto-optimal front. We numerically evaluate the performance of MOPBD* and its variants in various maps with two and three objectives. Results show that our approach is more efficient than search from scratch, and runs up to an order of magnitude faster than the existing incremental method for multi-objective path planning.

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。

↑