We present a novel approach to the verification of multi-agent systems using bounded model checking for specifications in LTLK, a linear time temporal-epistemic logic. The method is based on binary decision diagrams rather than the standard conversion to Boolean satisfiability. We apply the approach to two classes of interpreted systems: the standard, synchronous semantics and the interleaved semantics. We provide a symbolic algorithm for the verification of LTLK over models of multi-agent systems and evaluate its implementation against MCK, a competing model checker for knowledge. Our evaluation indicates that the interleaved semantics can often be preferable in the verification of LTLK.
展开▼
