An improved method for computing a generalized spectral coefficient

  • Authors:
  • D. M. Miller

  • Affiliations:
  • Dept. of Comput. Sci., Victoria Univ., BC

  • Venue:
  • IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems
  • Year:
  • 2006

Quantified Score

Hi-index 0.03

Visualization

Abstract

In 1995, Thornton and Nair studied the computation and application of a general class of spectral coefficients. A method for the computation of individual spectral coefficients using output probabilities computed on ROBDDs was presented. In this paper, that method is improved. Specifically, an alternative method for computing output probabilities is presented which, unlike the previous method, requires no recomputation for shared ROBDD representations and allows the use of edge negations. Based on this method, an alternative approach to computing the value of a spectral coefficient using output probabilities and ROBDDs is given. In contrast to the method described in 1995, one rather than two composition functions need to be considered, and no supplementary ROBDDs need to be constructed