论文信息 - Learning to Take Concurrent Actions

Learning to Take Concurrent Actions

We investigate a general semi-Markov Decision Process (SMDP) framework for modeling concurrent decision making, where agents learn optimal plans over concurrent temporally extended actions. We introduce three types of parallel termination schemes - all, any and continue - and theoretically and experimentally compare them.

Sridhar Mahadevan | Khashayar Rohanimanesh | S. Mahadevan | Khashayar Rohanimanesh

[1] Craig A. Knoblock. Generating Parallel Execution Plans with a Partial-order Planner , 1994, AIPS.

[2] Paweł Cichosz. Learning Multidimensional Control Actions From Delayed Reinforcements , 1995 .

[3] Raymond Reiter,et al. Natural Actions, Concurrency and Continuous Time in the Situation Calculus , 1996, KR.

[4] Satinder P. Singh,et al. How to Dynamically Merge Markov Decision Processes , 1997, NIPS.

[5] Ronen I. Brafman,et al. Planning with Concurrent Interacting Actions , 1997, AAAI/IAAI.

[6] Doina Precup,et al. Between MDPs and Semi-MDPs: A Framework for Temporal Abstraction in Reinforcement Learning , 1999, Artif. Intell..

[7] Sridhar Mahadevan,et al. Decision-Theoretic Planning with Concurrent Temporally Extended Actions , 2001, UAI.