def __solve_problem(self, hts: HTS, prop: Optional[FNode],
lemmas: Optional[List[FNode]],
assumptions: Optional[List[FNode]],
problem: NamedTuple) -> str:
trace = None
traces = None
region = None # only used for parametric model checking
accepted_ver = False
assert hts.assumptions is None, "There should not be any left-over assumptions from previous problems"
for assump in assumptions:
hts.add_assumption(assump)
for lemma in lemmas:
hts.add_lemma(lemma)
bmc_safety = BMCSafety(hts, problem)
bmc_parametric = BMCParametric(hts, problem)
bmc_ltl = BMCLTL(hts, problem)
res = VerificationStatus.UNC
bmc_length = problem.bmc_length
bmc_length_min = problem.bmc_length_min
if problem.verification == VerificationType.SAFETY:
accepted_ver = True
Logger.log("Property: %s" % (prop.serialize(threshold=100)), 2)
res, trace, _ = bmc_safety.safety(prop, bmc_length, bmc_length_min,
problem.processes)
if problem.verification == VerificationType.LTL:
accepted_ver = True
res, trace, _ = bmc_ltl.ltl(prop, bmc_length, bmc_length_min)
if problem.verification == VerificationType.SIMULATION:
accepted_ver = True
res, trace = bmc_safety.simulate(prop, bmc_length)
if problem.verification == VerificationType.PARAMETRIC:
accepted_ver = True
Logger.log("Property: %s" % (prop.serialize(threshold=100)), 2)
res, traces, region = bmc_parametric.parametric_safety(
prop,
bmc_length,
bmc_length_min,
ModelExtension.get_parameters(hts),
at_most=problem.cardinality)
if problem.verification == VerificationType.EQUIVALENCE:
accepted_ver = True
bmcseq = BMCSafety(hts, problem)
res, trace, t = bmcseq.safety(prop, bmc_length, bmc_length_min)
if not accepted_ver:
Logger.error("Invalid verification type")
Logger.log("\n*** Problem \"%s\" is %s ***" % (problem.name, res), 1)
return res, trace, traces, region
def combine_systems(hts,
hts2,
k,
symbolic_init,
eqprop=None,
inc=True,
non_deterministic=False):
htseq = HTS("eq")
hts1_varnames = [v.symbol_name() for v in hts.vars]
hts2_varnames = [v.symbol_name() for v in hts2.vars]
map1 = dict([(v, TS.get_prefix_name(v, S1)) for v in hts1_varnames]+\
[(TS.get_prime_name(v), TS.get_prefix_name(TS.get_prime_name(v), S1)) for v in hts1_varnames])
map2 = dict([(v, TS.get_prefix_name(v, S2)) for v in hts2_varnames]+\
[(TS.get_prime_name(v), TS.get_prefix_name(TS.get_prime_name(v), S2)) for v in hts2_varnames])
ts1_init = TRUE()
ts2_init = TRUE()
if not symbolic_init:
ts1_init = substitute(hts.single_init(), map1)
ts2_init = substitute(hts2.single_init(), map2)
ts1 = TS()
ts1.vars = set([TS.get_prefix(v, S1) for v in hts.vars])
ts1.set_behavior(ts1_init,\
substitute(hts.single_trans(), map1),\
substitute(hts.single_invar(), map1))
ts1.state_vars = set([TS.get_prefix(v, S1) for v in hts.state_vars])
ts2 = TS()
ts2.vars = set([TS.get_prefix(v, S2) for v in hts2.vars])
ts2.set_behavior(ts2_init,\
substitute(hts2.single_trans(), map2),\
substitute(hts2.single_invar(), map2))
ts2.state_vars = set([TS.get_prefix(v, S2) for v in hts2.state_vars])
htseq.add_ts(ts1)
htseq.add_ts(ts2)
assumptions = []
lemmas = []
def sets_intersect(set1, set2):
for el in set1:
if not el in set2:
return False
return True
if hts.assumptions is not None:
for assumption in hts.assumptions:
assumptions.append(assumption)
if hts.lemmas is not None:
for lemma in hts.lemmas:
lemmas.append(lemma)
if hts2.assumptions is not None:
for assumption in hts2.assumptions:
assumptions.append(assumption)
if hts2.lemmas is not None:
for lemma in hts2.lemmas:
lemmas.append(lemma)
for assumption in assumptions:
fv_assumption = get_free_variables(assumption)
c_assumption = TRUE()
if sets_intersect(fv_assumption, hts.vars):
c_assumption = And(c_assumption, substitute(assumption, map1))
if sets_intersect(fv_assumption, hts2.vars):
c_assumption = And(c_assumption, substitute(assumption, map2))
if c_assumption != TRUE():
htseq.add_assumption(c_assumption)
for lemma in lemmas:
fv_lemma = get_free_variables(lemma)
c_lemma = TRUE()
if sets_intersect(fv_lemma, hts.vars):
c_lemma = And(c_lemma, substitute(lemma, map1))
if sets_intersect(fv_lemma, hts2.vars):
c_lemma = And(c_lemma, substitute(lemma, map2))
if c_lemma != TRUE():
htseq.add_lemma(c_lemma)
miter_out = Symbol(EQS, BOOL)
inputs = hts.input_vars.intersection(hts2.input_vars)
outputs = hts.output_vars.intersection(hts2.output_vars)
htseq.input_vars = set([
TS.get_prefix(v, S1) for v in hts.input_vars
]).union(set([TS.get_prefix(v, S2) for v in hts2.input_vars]))
htseq.output_vars = set([
TS.get_prefix(v, S1) for v in hts.output_vars
]).union(set([TS.get_prefix(v, S2) for v in hts2.output_vars]))
if symbolic_init or (not non_deterministic):
states = hts.state_vars.intersection(hts2.state_vars)
else:
states = []
eqinputs = TRUE()
eqoutputs = TRUE()
eqstates = TRUE()
for inp in inputs:
eqinputs = And(
eqinputs,
EqualsOrIff(TS.get_prefix(inp, S1), TS.get_prefix(inp, S2)))
for out in outputs:
eqoutputs = And(
eqoutputs,
EqualsOrIff(TS.get_prefix(out, S1), TS.get_prefix(out, S2)))
for svar in states:
eqstates = And(
eqstates,
EqualsOrIff(TS.get_prefix(svar, S1), TS.get_prefix(svar, S2)))
if eqprop is None:
if symbolic_init or (not non_deterministic):
invar = And(eqinputs,
Iff(miter_out, Implies(eqstates, eqoutputs)))
else:
invar = And(eqinputs, Iff(miter_out, eqoutputs))
Logger.log('Inferring equivalence property: {}'.format(invar), 2)
else:
sparser = StringParser()
eqprop = sparser.parse_formulae(eqprop)
if len(eqprop) > 1:
Logger.error("Expecting a single equivalence property")
eqprop = eqprop[0][1]
invar = Iff(miter_out, eqprop)
Logger.log('Using provided equivalence property: {}'.format(invar),
2)
tsmo = TS()
tsmo.vars = set([miter_out])
tsmo.invar = invar
htseq.add_ts(tsmo)
return (htseq, miter_out)
