The modeling steps needed to create dynamically based automated contro
l of tokamak plasmas are traced. This involves integrating models of c
urrent/magnetic-flux dynamics; plasma transport; plasma geometry; and
source terms, such as lower hybrid, fast wave, and pellet-fueling depo
sition. Perturbative analysis of these models then yields the linear r
esponse of the tokamak to changes in coil voltages, applied radio-freq
uency power,; or pellet-firing frequency. Comparison of the linear res
ponse models to nonlinear numerical calculations reveals that the plas
ma position and shape modeling will require future refinements.