arXivDaily arXiv每日学术速递 周一至周五更新
北美退伍军人节(Veterans Day),12 日无更新
arXiv 2405.02456math.OCcs.LG
收藏

用于约束多任务强化学习的自然策略梯度与演员-评论家方法

Natural Policy Gradient and Actor Critic Methods for Constrained Multi-Task Reinforcement Learning

JPMorgan AI Research(摩根大通人工智能研究院);Virginia Tech(弗吉尼亚理工大学);Georgia Tech(佐治亚理工学院)

Sihan Zeng, Thinh T. Doan, Justin Romberg

AI总结:

本文针对约束多任务强化学习提出集中式与分散式求解框架,设计原始-对偶算法及基于采样的演员-评论家算法,并扩展到线性函数逼近,以实现满足单任务性能约束的全局最优策略。

摘要:

多任务强化学习(RL)旨在找到一个单一策略,能够同时有效地解决多个任务。本文提出了一种用于多任务RL的约束形式化,其目标是在每个任务性能有界约束的前提下,最大化策略在所有任务上的平均性能。我们考虑在集中式设置和分散式设置中解决该问题:在集中式设置中,所有任务的信息可由单个服务器访问;在分散式设置中,一个智能体网络中的每个智能体被赋予一个任务并观察局部信息,通过局部通信协作找到全局约束目标的解。我们首先提出一种原始-对偶算法,证明在精确梯度评估下该算法可收敛到该约束形式的全局最优解。当梯度未知时,我们进一步开发了一种基于采样的演员-评论家算法,利用状态、动作和奖励的在线样本找到最优策略。最后,我们研究了该算法在线性函数逼近设置下的扩展。

英文摘要:

Multi-task reinforcement learning (RL) aims to find a single policy that effectively solves multiple tasks at the same time. This paper presents a constrained formulation for multi-task RL where the goal is to maximize the average performance of the policy across tasks subject to bounds on the performance in each task. We consider solving this problem both in the centralized setting, where information for all tasks is accessible to a single server, and in the decentralized setting, where a network of agents, each given one task and observing local information, cooperate to find the solution of the globally constrained objective using local communication. We first propose a primal-dual algorithm that provably converges to the globally optimal solution of this constrained formulation under exact gradient evaluations. When the gradient is unknown, we further develop a sampled-based actor-critic algorithm that finds the optimal policy using online samples of state, action, and reward. Finally, we study the extension of the algorithm to the linear function approximation setting.

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。

↑