On Optimal Satisficing: How Simple Policies Can Achieve Excellent Results

Abstract

Herbert Simon introduced the notion of satisficing to explain how boundedly rational agents might approach difficult sequential decision problems. His satisficing decision makers were offered as an alternative to optimizers, who need impressive computational capacities in order to maximize their payoffs. In this chapter, we present a simplified sequential search problem for a satisficing decision maker, and show how to compute its optimal satisficing search policies. Our analyses provide bounds on the performance of satisficing search policies.