A space-efficient on-the-fly algorithm for real-time model checking

Abstract

In temporal-logic model checking, we verify the correctness of a program with respect to a desired behavior by checking whether a structure that models the program satisfies a temporal-logic formula that specifies the behavior. The main practical limitation of model checking is caused by the size of the state space of the program, which grows exponentially with the number of concurrent components. This problem, known as the state-explosion problem, becomes more difficult when we consider real-time model checking, where the program and the specification involve quantitative references to time. In particular, when use timed automata to describe real-time programs and we specify timed behaviors in the logic TCTL, a real-time extension of the temporal logic CTL with clock variables, then the state space under consideration grows exponentially not only with the number of concurrent components, but also with the number of clocks and the length of the clock constraints used in the program and the specification. Two powerful methods for coping with the state-explosion problem are on-the-fly and space-efficient model checking. In on-the-fly model checking, we explore only the portion of the state space of the program whose exploration is essential for determining the satisfaction of the specification. In space-efficient model checking, we store in memory the minimal information required, preferring to spend time on reconstructing information rather than spend space on storing it. In this work we develop an automata-theoretic approach to TCTL model checking that combines both methods. We suggest, for the first time, a PSPACE on-the-fly model-checking algorithm for TCTL.

Supported in part by the ONR YIP award N00014-95-1-0520, by the NSF CAREER award CCR-9501708, by the NSF grant CCR-9504469, by the AFOSR contract F49620-93-1-0056, and by the ARPA grant NAG2-892.