Efficient Multiple-Valued Model-Checking Using Lattice Representations

Abstract

Multiple-valued logics can be effectively used to reason about incomplete and/or inconsistent systems, e.g. during early software requirements or as the systems evolve. We specify multiple-valued logics using finite lattices. In this paper, we use lattice representation theory to cast the multiple-valued model-checking problem in terms of symbolic operations on classical sets of states, provided the lattices are distributive. This allows us to partially reuse existing symbolic model-checking technology and improve efficiency over previous implementations that were based on multiple-valued decision diagrams.