William G. Kennedy and Jeffrey K. Bassett

Large-scale social simulations require a cognitively credible but computationally efficient cognitive architecture to support simulations with thousands to tens of thousands agents. In previous work developing and experimenting with a large-scale social simulation, we successfully employed an ad hoc cognitive model, a formula-based decision function. We explain why we now believe that a “fast and frugal” cognitive architecture to be superior based on its indistinguishable computational efficiency and much better cognitive plausibility.

Download PDF