Re: fairness constraint

From: Fabio Somenzi (Fabio@colorado.edu)
Date: Thu May 24 2001 - 13:09:28 MDT

  • Next message: Roderick Bloem: "Counterexamples and inputs"

    A fairness constraint "a=1" instructs the model checker to disregard
    computations (paths) along which "a=1" does not happen infinitely
    often. It is not necessary for "a=1" to always hold along the path
    for it to be fair.

    Hence, it is quite possible for a counterexample to contain states
    with "a=0" even if the fairness constraint "a=1" is in effect.

    >From a practical standpoint, it's worth noting that in Vis, fairness
    constraints are honored by "model_check," and ignored by
    "check_invariant."

    It is unusual to check invariants under fairness constraints, which
    are most often specified for liveness properties. However, should one
    want to check whether an invariant holds in all states reachable along
    a fair path (as opposed to the states reachable via any path), the
    invariant should be written as an AG property, and checked with "mc."

    >>>>> "vg" == vanessa goyal <vanessa_goyal@rediffmail.com> writes:

     vg> sir
     vg> If a=1(one of the input to FSM) is fairness constraint,then
     vg> reachability analysis will get restricted to paths for which a=1.
     vg> So,i expect, on failure of any property,counter example generated
     vg> should contain only those states for which a=1,but its showing
     vg> states having a=0 also.
     vg> Is it my misconception.
     vg> thanking you
     vg> vanessa

    Fabio



    This archive was generated by hypermail 2b30 : Thu May 24 2001 - 13:14:23 MDT