Efficient Abstract Interpretation using Component-Wise Homomorphism
Jörg Köller and Markus Mohnen
To appear at the 2nd International Conference on Principles and Practice of
Declarative Programming (PPDP 2000), Montreal, Canada, September 20-22, 2000
Abstract
Fixed point based abstract interpretation requires compact
representations of functions. In general, finding a fixed point
requires exponential time. Good representations allow the reduction of
the fixed point solvers computational complexity. One approach for
obtaining compact representations is the use of special classes of
functions. In this paper, we focus on the class of component-wise
homomorphisms. We show that this class of functions gives more compact
representations than component-wise additive functions.