arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2202.07960cs.LGcs.AImath.APmath.OC
收藏

随机环境下连续时间与连续状态的时序差分学习

Temporal Difference Learning with Continuous Time and State in the Stochastic Setting

Inria(法国国家信息与自动化研究所);École Normale Supérieure(巴黎高等师范学院);Institut Louis Bachelier(路易·巴舍利耶研究所);PSL Research University(巴黎文理研究大学)

Ziad Kobeissi, Francis Bach

AI总结:

针对连续时间策略评估,提出无模型与基于模型的 TD(0) 变体,证明收敛速率并经仿真验证,还可用于求解线性 PDE 或 BSDE。

摘要:

我们研究连续时间策略评估问题,即通过观测学习与无控制的连续时间随机动态及奖励函数相关联的价值函数。我们提出著名 TD(0) 方法的两种采用趋于零时间步长的原创变体:一种无模型,另一种基于模型。对于这两种方法,我们证明了理论收敛速率,并随后通过数值仿真加以验证。此外,这些方法也可解释为用于逼近线性 PDE(偏微分方程)或线性 BSDE(倒向随机微分方程)解的新型强化学习方法。

英文摘要:

We consider the problem of continuous-time policy evaluation. This consists in learning through observations the value function associated with an uncontrolled continuous-time stochastic dynamic and a reward function. We propose two original variants of the well-known TD(0) method using vanishing time steps. One is model-free and the other is model-based. For both methods, we prove theoretical convergence rates that we subsequently verify through numerical simulations. Alternatively, those methods can be interpreted as novel reinforcement learning approaches for approximating solutions of linear PDEs (partial differential equations) or linear BSDEs (backward stochastic differential equations).

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。

↑