Deciding a Language of Sets (and Relations)
Language
Consider a simple language of sets, motivated by verification of programs that manipulate containers:
Question: how do we justify ?
Idea of Decidability
We show that this language is decidable by reduction to universal class.
Convert formula to negation-normal form.
Express all operations using quantifiers.
Note: the resulting formula has a bounded number of universal quantifiers
- this gives good complexity of the generated formulas (in NP if the constants are written in unary)