State Compression in SPIN: Recursive Indexing and Compression Training Runs

The verification algorithm of SPIN is based on an explicit enumeration of a subset of the reachable state-space of a system that is obtained through the formalization of a correctness requirement as an ω-automaton. This ω-automaton restricts the state-space to precisely the subset that may contain the counter-examples to the original correctness requirement, if they exist. This method of verification conforms to the method for automata-theoretic verification outlined in [VW86]. SPIN derives much of its power from an efficient implementation of the explicit state enumeration method in this context. The maximum number of reachable states that can be stored in main memory with this method, either explicitly or implicitly, then determines the maximum problem size that can be solved. We review some methods that have been tried to reduce the amount memory used per state, and describe some new methods that achieve relatively high compression rates. Index Terms — Model Checking, Verification, State Compression.