Modular Primal-Dual Fixpoint Logic Solving for Temporal Verification
summary
Reduces μCLP validity to a predicate-constraint problem and leverages an existing solver to give a sound and relatively-complete decision method, then improves convergence by running modular primal-dual solves in parallel that exchange partial solutions as sound upper/lower bounds (exploiting μCLP’s closure under complement). Encodings map loops and subformulas to per-predicate (co)inductive definitions to capture termination, non-termination, LTL, CTL and full modal μ-calculus on infinite-state programs; implementation on benchmarks evaluates with promising results.
AI-written summary · about