An Approximate Solution Method for Large Risk-Averse Markov Decision Processes

Abstract

Stochastic domains often involve risk-averse decision makers. While recent work has focused on how to model risk in Markov decision processes using risk measures, it has not addressed the problem of solving large risk-averse formulations. In this paper, we propose and analyze a new method for solving large risk-averse MDPs with hybrid continuous-discrete state spaces and continuous action spaces. The proposed method iteratively improves a bound on the value function using a linearity structure of the MDP. We demonstrate the utility and properties of the method on a portfolio optimization problem.

Cite

Text

Petrik and Subramanian. "An Approximate Solution Method for Large Risk-Averse Markov Decision Processes." Conference on Uncertainty in Artificial Intelligence, 2012.

Markdown

[Petrik and Subramanian. "An Approximate Solution Method for Large Risk-Averse Markov Decision Processes." Conference on Uncertainty in Artificial Intelligence, 2012.](https://mlanthology.org/uai/2012/petrik2012uai-approximate/)

BibTeX

@inproceedings{petrik2012uai-approximate,
  title     = {{An Approximate Solution Method for Large Risk-Averse Markov Decision Processes}},
  author    = {Petrik, Marek and Subramanian, Dharmashankar},
  booktitle = {Conference on Uncertainty in Artificial Intelligence},
  year      = {2012},
  pages     = {805-814},
  url       = {https://mlanthology.org/uai/2012/petrik2012uai-approximate/}
}