Making value-laden decisions under conflict

Optimal decision-making is, prima facie, the first preference of any agent tasked with making choices. But if the demand for optimality requires the decision problem to be modified in an arbitrary and possibly objectionable way, then the effort to obtain an optimal solution may be misplaced. An alternative is to formulate satisficing decisions based upon local, rather than global, performance criteria. Epistemic utility theory provides a criteria-based satisficing decision-making procedure. An essential component of this theory is the notion that agents often make value-laden decisions, and that the informational value of decisions must be explicitly modeled. The assignment of value engenders conflict, however, since decisions may be made, based not only on their credibility, but also on the value ascribed to them.<<ETX>>