ABSTRACT
This chapter expands on how resources are assigned to matching production-rule instantiations, how production-rule instantiations are evaluated, and how people determine that an instantiation satisfices. In the current context, with goals explicitly mentioned in the production rules, optimization means maximizing the probability of achieving these goals while minimizing the cost of doing so or, to put it more precisely, maximizing the expected utility, where this is defined as expected gain minus expected cost. It presents a rational analysis of how production strength and chunk activation should combine to determine the time to match a chunk to a production. It discusses the effects of production-rule strength. Anderson and Schooler presented an analysis of why time to retrieve a memory trace might be a power function of odds. Because cognition must be implemented in the human brain, it seems transparent that the implementation of ACT-R should be in terms of neural-like computations.
