Sciweavers

Free Online Productivity Tools i2Speak i2Symbol i2OCR iTex2Img iWeb2Print iWeb2Shot i2Type iPdf2Split iPdf2Merge i2Bopomofo i2Arabic i2Style i2Image i2PDF iLatex2Rtf Sci2ools

25

COLT
2004
Springer

favoriteEmaildiscussreport

99views Machine Learning» more COLT 2004»

Reinforcement Learning for Average Reward Zero-Sum Games

14 years 4 months ago

Reinforcement Learning for Average Reward Zero-Sum Games

Download www.ece.mcgill.ca

Abstract. We consider Reinforcement Learning for average reward zerosum stochastic games. We present and analyze two algorithms. The ﬁrst is based on relative Q-learning and the second on Q-learning for stochastic shortest path games. Convergence is proved using the ODE (Ordinary Diﬀerential Equation) method. We further discuss the case where not all the actions are played by the opponent with comparable frequencies and present an algorithm that converges to the optimal Q-function, given the observed play of the opponent.

Shie Mannor

Real-time Traffic

COLT 2004 | Ordinary Diﬀerential Equation | Stochastic Shortest Path | Zerosum Stochastic Games |

claim paper

Related Content

» The Steering Approach for MultiCriteria Reinforcement Learning

» RMAX A General Polynomial Time Algorithm for NearOptimal Reinforcement Learning

» Formalizing Multistate Learning Dynamics

» Scaling ModelBased AverageReward Reinforcement Learning for Product Delivery

» ModelBased Average Reward Reinforcement Learning

» Hierarchically Optimal Average Reward Reinforcement Learning

» Sensitive Discount Optimality Unifying Discounted and Average Reward Reinforcement Learnin...

» Sparse reward processes

» ContinuousTime Hierarchical Reinforcement Learning

Post Info
More Details (n/a)

Added	01 Jul 2010
Updated	01 Jul 2010
Type	Conference
Year	2004
Where	COLT
Authors	Shie Mannor

Comments (0)