Minimax regret based elicitation of generalized additive utilities
Artificial Intelligence
2012-06-26 v1
Abstract
We describe the semantic foundations for elicitation of generalized additively independent (GAI) utilities using the minimax regret criterion, and propose several new query types and strategies for this purpose. Computational feasibility is obtained by exploiting the local GAI structure in the model. Our results provide a practical approach for implementing preference-based constrained configuration optimization as well as effective search in multiattribute product databases.
Keywords
Cite
@article{arxiv.1206.5255,
title = {Minimax regret based elicitation of generalized additive utilities},
author = {Darius Braziunas and Craig Boutilier},
journal= {arXiv preprint arXiv:1206.5255},
year = {2012}
}
Comments
Appears in Proceedings of the Twenty-Third Conference on Uncertainty in Artificial Intelligence (UAI2007)