Conditional Utility, Utility Independence, and Utility Networks
This work addresses foundational issues in decision theory and AI for researchers and practitioners, but it appears incremental as it builds on prior concepts like utility distributions.
The paper tackles the problem of representing utility functions by introducing a new interpretation of conditional utility and utility independence, showing that utility distributions can be viewed as a special case of additive multiattribute utility functions and presenting utility networks for compact representation.
We introduce a new interpretation of two related notions - conditional utility and utility independence. Unlike the traditional interpretation, the new interpretation renders the notions the direct analogues of their probabilistic counterparts. To capture these notions formally, we appeal to the notion of utility distribution, introduced in previous paper. We show that utility distributions, which have a structure that is identical to that of probability distributions, can be viewed as a special case of an additive multiattribute utility functions, and show how this special case permits us to capture the novel senses of conditional utility and utility independence. Finally, we present the notion of utility networks, which do for utilities what Bayesian networks do for probabilities. Specifically, utility networks exploit the new interpretation of conditional utility and utility independence to compactly represent a utility distribution.