一句话：研究 MDP 中基于分布的目标如何在完全信息与部分信息之间取得策略层面的权衡。
论文研究马尔可夫决策过程（MDP）中基于分布的目标。这类目标在完全信息与部分信息之间形成一种有趣权衡：与完全可观察情形类似，策略可以依赖系统状态；但与部分可观察情形类似，策略又必须同时兼顾所有状态。作者围绕这一目标类别展开形式化分析，关注其策略结构与信息约束，揭示完全信息和部分信息设定之间的张力。摘要未给出具体算法或实验评估，因此贡献主要体现在对该类目标的概念刻画与理论定位。
We consider distribution-based objectives for Markov Decision Processes (MDP). This class of objectives gives rise to an interesting trade-off between full and partial information. As in full observation, the strategy in the MDP can depend on the state of the system, but similar to partial information, the strategy needs to account for all the states at the same time.
DOI 原文 ·
@inproceedings{paperbot4338,
title = {Distribution-based objectives for Markov Decision Processes},
author = {S. Akshay and Blaise Genest and Nikhil Vyas},
booktitle = {LICS},
year = {2018},
doi = {10.1145/3209108.3209185}
}