Lazy group sifting for efficient symbolic state traversal of FSMs

Proposes lazy group sifting for dynamic variable reordering during state traversal of finite state machines (FSMs). The proposed method relaxes the idea of pairwise grouping of the present state variables and their corresponding next state variables. This is done to produce better variable orderings during image computation without causing BDD (binary decision diagram) size blowup in the substitution of next state variables with present state variables at the end of image computation. Experimental results show that our approach is more robust in state traversal than the approaches that either unconditionally group variable pairs or never group them.