Logic Programming , Boolean using μ - calculus Model Checking on a Boolean Domain

computation. BDDs have been very successful in this Abstract respect. One of the main obstacles to automatic verification A global optimization is a transformation taking an of Finite State Systems (FSSs) is state explosion. In MC problem (M, φ) and returning an (hopefully) easier this respect automatic verification of an FSS M using MC problem (M', φ') s.t. answer(M, φ) = answer(M', φ'). Model Checking and Binary Decision Diagrams (BDDs) E. g. in [2] sec. 5 and [11] are optimization techniques in has an intrinsic limitation: no automatic global which φ (but not M) is modified to improve fixpoint optimization of the verification task is possible until a computation performances. All MC optimization BDD representation for M is generated. This is because techniques that we know of act only on φ. However to systems and specifications are defined using different avoid state explosion when dealing with combinatorial languages. To perform global optimization before circuits we need to modify M and φ. This is because generating a BDD representation for M we propose to use BDDs are a canonical form for boolean functions. To the the same language to define systems and specifications. best of our knowledge no automatic global (i.e. acting on We show that First Order Logic on a Boolean both M and φ) optimization technique has been presented Domain yields an efficient functional programming in the literature. language that can be used to represent, specify and Automatic global optimization in an MC setting is automatically verify FSSs. E.g. on a SUN Sparc Station difficult because model M and specification φ are defined 2 we were able to automatically verify a 64 bit using different languages. E.g. M can be defined using commercial multiplier. Hardware Description Languages, Process Algebras, etc.,