arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1207.4150cs.AI
收藏

用连续变量和离散变量求解因子化的MDP

Solving Factored MDPs with Continuous and Discrete Variables

Berkeley Research Center(伯克利研究中心);Intel Corporation(英特尔公司);University of Pittsburgh(匹兹堡大学)

Carlos E. Guestrin, Milos Hauskrecht, Branislav Kveton

摘要:

虽然许多现实世界的随机规划问题更自然地制定了离散和连续变量的混合模型,目前的最先进的方法不能充分解决这些问题。我们提出了第一个框架,可以利用问题结构建模和有效地解决混合问题。我们制定这些问题的混合马尔可夫决策过程(连续和离散的状态和动作变量的MDPs),我们假设可以表示在一个因素的方式使用混合动态贝叶斯网络(混合DBN)。这种提法也使我们能够将我们的方法应用于协作多智能体设置。我们提出了一种新的线性规划逼近方法,利用混合MDP的结构,让我们更有效地计算近似价值函数。特别是,我们描述了一个新的连续变量的分解离散化,避免了传统方法的指数爆炸。我们提供了理论界的质量,这样的近似和它的规模扩大的潜力。我们支持我们的理论论点与一组控制问题的实验高达28维连续状态空间和22维动作空间。

英文摘要:

Although many real-world stochastic planning problems are more naturally formulated by hybrid models with both discrete and continuous variables, current state-of-the-art methods cannot adequately address these problems. We present the first framework that can exploit problem structure for modeling and solving hybrid problems efficiently. We formulate these problems as hybrid Markov decision processes (MDPs with continuous and discrete state and action variables), which we assume can be represented in a factored way using a hybrid dynamic Bayesian network (hybrid DBN). This formulation also allows us to apply our methods to collaborative multiagent settings. We present a new linear program approximation method that exploits the structure of the hybrid MDP and lets us compute approximate value functions more efficiently. In particular, we describe a new factored discretization of continuous variables that avoids the exponential blow-up of traditional approaches. We provide theoretical bounds on the quality of such an approximation and on its scale-up potential. We support our theoretical arguments with experiments on a set of control problems with up to 28-dimensional continuous state space and 22-dimensional action space.

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。

↑