SANLab-CM – The Stochastic Activity Network Laboratory for Cognitive Modeling

SANLab-CM is an activity network tool for the stochastic modeling of routine interactive behavior. Within the cognitive engineering community the best-known examples of activity networking modeling are the CPM-GOMS models of Project Ernestine (Gray, John, & Atwood, 1993). Project Ernestine showed that modeling the parallel use of cognitive, perceptual, and motor resources within an activity network formalism produces reliable and accurate predictions of expert performance times across alternative designs for the same task. SANLab-CM provides time predictions, but its essence is the prediction of procedural variability amidst strategic constancy: when expert human performers follow the same task strategy from trial to trial variability in the processing time of cognitive, perceptual, and motor resources is such as to produce different critical paths of performance and significantly different execution times. The stochastic component of SANLab-CM goes beyond current techniques to create a new means of assessing alternative designs based on the procedural variability expected in expert performance.