Loading...
Please wait, while we are loading the content...
Similar Documents
The Steering Approach for Multi-Criteria Reinforcement (2002)
| Content Provider | CiteSeerX |
|---|---|
| Author | Mannor, Shie Shimkin, Nahum |
| Description | We consider the problem of learning to attain multiple goals in a dynamic environment, which is initially unknown. In addition, the environment may contain arbitrarily varying elements related to actions of other agents or to non-stationary moves of Nature. This problem is modelled as a stochastic (Markov) game between the learning agent and an arbitrary player, with a vector-valued reward function. The objective of the learning agent is to have its long-term average reward vector belong to a given target set. We devise an algorithm for achieving this task, which is based on the theory of approachability for stochastic games. This algorithm combines, in an appropriate way, a finite set of standard, scalar-reward learning algorithms. Sufficient conditions are given for the convergence of the learning algorithm to a general target set. The specialization of these results to the single-controller Markov decision problem are discussed as well. 1 |
| File Format | |
| Language | English |
| Publisher Date | 2002-01-01 |
| Publisher Institution | Learning, Advances in Neural Information Processing Systems |
| Access Restriction | Open |
| Subject Keyword | Algorithm Combine Stochastic Game Appropriate Way Sufficient Condition Steering Approach Single-controller Markov Decision Problem Arbitrary Player Learning Agent Long-term Average Reward Vector Belong Multiple Goal Multi-criteria Reinforcement Scalar-reward Learning Algorithm Non-stationary Move Dynamic Environment Finite Set Vector-valued Reward Function General Target |
| Content Type | Text |
| Resource Type | Article |
