Share on Facebook Tweet on Twitter Share on LinkedIn Share by email
Generalized, Efficient Array Decision Procedures

Leonardo de Moura and Nikolaj Bjorner

Abstract

The theory of arrays is ubiquitous in the context of automatic software and hardware verification and symbolic analysis. The basic array theory was introduced by McCarthy and allows to symbolically representing array updates.

To this date the theory of arrays itself remains fundamental to how modern program verification, test-case generation, and model-based program tools model program heaps and higher level data-types, such as sets and finite maps.

We present combinatory array logic, CAL, using a small, but powerful core of combinators, and reduce it to the theory of uninterpreted functions. CAL allows expressing properties that go well beyond the basic array theory. CAL does not allow expressing the identity function I. CAL+I on the other hand allows encoding arbitrary lambda terms. We provide a new efficient decision procedure for the base theory as well as CAL. The efficient procedure serves a critical role in the performance of the state-of-the-art SMT solver Z3 on array formulas from applications.

Details

Publication typeTechReport
NumberMSR-TR-2009-121
PublisherIEEE
> Publications > Generalized, Efficient Array Decision Procedures