Parametric analog behavioral modeling based on cancellation-free DDDs

This paper presents an efficient approach to generating simplified symbolic expressions for behavioral modeling of large linear analog circuits. The approach is based on a compact determining decision diagram representation of exact transfer functions and characteristics of analog circuits. We show how cancellation-free, s-expended DDDs (sDDDs) can be reconstructed from DDDs. With several DDD properties characterized in this paper, we compare two efficient algorithms for finding dominant terms based respectively on shortest paths and dynamic programming. We show that all these algorithms take time linear in terms of the number of DDD vertices. Experimental results show that our shortest path based algorithm is more memory efficient than the dynamic programming based algorithm.