A causal approach to hierarchical decomposition of factored MDPs

A causal approach to hierarchical decomposition of factored MDPs
复制标题

因式 MDP 层次分解的因果方法

DOI:
--
复制
发表时间:
2005
期刊:
International Conference on Machine Learning
影响因子:
--
通讯作者:
A. Barto
A. Barto
中科院分区:
--
文献类型:
--
作者:
Anders Jonsson;A. Barto

文献摘要

被引文献

相似文献

我们提出了可变影响结构分析,一个算法,动态地执行层次分解的因素马尔可夫决策过程。我们的算法确定状态变量之间的因果关系,并引入时间扩展的动作,导致状态变量的值发生变化。每一个时间上扩展的动作对应的子任务比整个任务更容易解决。实验结果表明,在扩展到更大的任务有很大的希望。
We present Variable Influence Structure Analysis, an algorithm that dynamically performs hierarchical decomposition of factored Markov decision processes. Our algorithm determines causal relationships between state variables and introduces temporally-extended actions that cause the values of state variables to change. Each temporally-extended action corresponds to a subtask that is significantly easier to solve than the overall task. Results from experiments show great promise in scaling to larger tasks.