Bounded Rationality: Models of Fast and Frugal Inference

I specify general criteria for models of bounded rationality and discuss specific models for satisficing inference. The task of these fast and frugal algorithms is to infer unknown features of their environment under the constraints of limited knowledge, limited time, and limited computational capacities. These algorithms violate fundamental tenets of classical rationality: They neither look up nor integrate all information. I review the performance of the satisficing "Take The Best" algorithm. Despite its frugality, Take The Best can make as many correct inferences as computationally expensive weighted linear models that use and combine all available information. Accurate inferences need not follow the dictates of classical rationality.