Abstract

An agent in a decision problem may not know the goals that should guide selection of an option. Accommodating this ignorance require methods that supplement expected utility theory.