Spectral decision diagrams using graph transformations
Mitchell Aaron Thornton, Rolf Drechsler · 2001
Spectral techniques are powerful methods for synthesis and verification of digital circuits. The advances in DD representations for discrete valued functions in terms of computational efficiency can be exploited in the calculation of the spectra of Boolean functions. The classical approach in computing the spectrum of a function by taking advantage of factored transformation matrices as used in the "Fast Fourier Transform" may be reformulated in terms of DD based graph algorithms resulting in a complete representation of the spectrum. The relationship between DD based interpretations and the linear algebra based definitions of spectral methods are described. 1 Introduction The proliferation of the use of spectral techniques for signal analysis and linear systems analysis and design motivates researchers to look for ways to apply these methods to digital systems. The pioneering work of Karpovsky [14] and Lechner [15] are generally regarded as the basis of subsequent work in this field...