A probabilistic method to estimate the switching activity of a combinationa
l circuit under a real-gate delay model considering temporal, structural an
d input pattern dependencies is introduced. It is proved that the switching
activity evaluation problem is reduced to the zero-delay problem at specif
ic time instances. A mathematical model based on Markov stochastic processe
s, which describes the temporal and spatial correlation in terms of the ass
ociated zero-delay parameters, is presented. To handle the influence of tim
e on glitch generation, the theory of the timed Boolean function (TBF) is a
dopted. Additionally, an algorithm to evaluate the switching activity at sp
ecific time instances using TBF-ordered binary decision diagrams (TBF-OBDDs
) is given. Comparative study of benchmark circuits demonstrates the accura
cy and efficiency of the proposed method.