arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 1909.12499cs.ROcs.AImath.OC
收藏

不确定条件下的风险规避规划

Risk-Averse Planning Under Uncertainty

California Institute of Technology(加州理工学院);NASA Jet Propulsion Laboratory(美国国家航空航天局喷气推进实验室)

Mohamadreza Ahmadi, Masahiro Ono, Michel D. Ingham, Richard M. Murray, Aaron D. Ames

摘要:

研究了具有动态一致风险目标的部分可观测马尔可夫决策过程的策略设计问题。为POMDP合成风险规避最优策略需要无限的内存,因此是不可判定的。为了克服这个困难,我们提出了一种基于有界策略迭代的方法来设计随机但有限状态(记忆)控制器,它利用了标准的凸优化方法。给定一个内存预算和最优性准则,该方法修改随机有限状态控制器,导致次优解具有较低的相干风险。

英文摘要:

We consider the problem of designing policies for partially observable Markov decision processes (POMDPs) with dynamic coherent risk objectives. Synthesizing risk-averse optimal policies for POMDPs requires infinite memory and thus undecidable. To overcome this difficulty, we propose a method based on bounded policy iteration for designing stochastic but finite state (memory) controllers, which takes advantage of standard convex optimization methods. Given a memory budget and optimality criterion, the proposed method modifies the stochastic finite state controller leading to sub-optimal solutions with lower coherent risk.

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。

↑