Flexible Policy Construction by Information Refinement

Abstract

We report on work towards flexible algorithms for solving decision problems represented as influence diagrams. An algorithm is given to construct a tree structure for each decision node in an influence diagram. Each tree represents a decision function and is constructed incrementally. The improvements to the tree converge to the optimal decision function (neglecting computational costs) and the asymptotic behaviour is only a constant factor worse than dynamic programming techniques, counting the number of Bayesian network queries. Empirical results show how expected utility increases with the size of the tree and the number of Bayesian net calculations.

Cite

Text

Horsch and Poole. "Flexible Policy Construction by Information Refinement." Conference on Uncertainty in Artificial Intelligence, 1996. doi:10.14288/1.0051145

Markdown

[Horsch and Poole. "Flexible Policy Construction by Information Refinement." Conference on Uncertainty in Artificial Intelligence, 1996.](https://mlanthology.org/uai/1996/horsch1996uai-flexible/) doi:10.14288/1.0051145

BibTeX

@inproceedings{horsch1996uai-flexible,
  title     = {{Flexible Policy Construction by Information Refinement}},
  author    = {Horsch, Michael C. and Poole, David L.},
  booktitle = {Conference on Uncertainty in Artificial Intelligence},
  year      = {1996},
  pages     = {315-324},
  doi       = {10.14288/1.0051145},
  url       = {https://mlanthology.org/uai/1996/horsch1996uai-flexible/}
}