3
0
Fork 0
mirror of https://github.com/Z3Prover/z3 synced 2025-06-06 14:13:23 +00:00

push blocking code to optimizer context

Signed-off-by: Nikolaj Bjorner <nbjorner@microsoft.com>
This commit is contained in:
Nikolaj Bjorner 2013-10-29 20:26:54 -07:00
parent b0fddd8e60
commit bc44bcad10
11 changed files with 152 additions and 49 deletions

View file

@ -46,83 +46,92 @@ Notes:
namespace opt {
class scoped_push {
opt_solver& s;
public:
scoped_push(opt_solver& s):s(s) {
s.push();
}
~scoped_push() {
s.pop(1);
}
};
void optimize_objectives::set_cancel(bool f) {
m_cancel = true;
}
void optimize_objectives::set_max(vector<inf_eps>& dst, vector<inf_eps> const& src) {
for (unsigned i = 0; i < src.size(); ++i) {
if (src[i] > dst[i]) {
dst[i] = src[i];
}
}
}
/*
Enumerate locally optimal assignments until fixedpoint.
*/
lbool mathsat_style_opt(
opt_solver& s,
app_ref_vector const& objectives,
vector<inf_eps_rational<inf_rational> >& values)
lbool optimize_objectives::basic_opt(app_ref_vector& objectives, vector<inf_eps>& values)
{
ast_manager& m = objectives.get_manager();
arith_util autil(m);
s.reset_objectives();
values.reset();
// First check_sat call to initialize theories
lbool is_sat = s.check_sat(0, 0);
if (is_sat == l_false) {
return is_sat;
}
s.push();
scoped_push _push(s);
opt_solver::toggle_objective _t(s, true);
for (unsigned i = 0; i < objectives.size(); ++i) {
s.add_objective(objectives[i]);
s.add_objective(objectives[i].get());
values.push_back(inf_eps(rational(-1),inf_rational(0)));
}
is_sat = s.check_sat(0, 0);
while (is_sat == l_true) {
// Extract values for objectives
values.reset();
values.append(s.get_objective_values());
while (is_sat == l_true && !m_cancel) {
set_max(values, s.get_objective_values());
IF_VERBOSE(1,
for (unsigned i = 0; i < values.size(); ++i) {
verbose_stream() << values[i] << " ";
}
verbose_stream() << "\n";);
expr_ref_vector disj(m);
expr_ref constraint(m), num(m);
for (unsigned i = 0; i < objectives.size(); ++i) {
expr_ref constraint(m);
if (!values[i].get_infinity().is_zero()) {
continue;
}
num = autil.mk_numeral(values[i].get_rational(), m.get_sort(objectives[i]));
SASSERT(values[i].get_infinitesimal().is_nonpos());
if (values[i].get_infinitesimal().is_neg()) {
disj.push_back(autil.mk_ge(objectives[i], num));
}
else {
disj.push_back(autil.mk_gt(objectives[i], num));
}
for (unsigned i = 0; i < objectives.size(); ++i) {
inf_eps const& v = values[i];
disj.push_back(s.block_lower_bound(i, v));
}
constraint = m.mk_or(disj.size(), disj.c_ptr());
s.assert_expr(constraint);
is_sat = s.check_sat(0, 0);
}
s.pop(1);
if (is_sat == l_undef) {
return is_sat;
if (m_cancel || is_sat == l_undef) {
return l_undef;
}
return l_true;
return l_true;
}
/**
Takes solver with hard constraints added.
Returns an optimal assignment to objective functions.
*/
lbool optimize_objectives(opt_solver& s,
app_ref_vector& objectives,
vector<inf_eps_rational<inf_rational> >& values) {
return mathsat_style_opt(s, objectives, values);
lbool optimize_objectives::operator()(app_ref_vector& objectives, vector<inf_eps>& values) {
return basic_opt(objectives, values);
}
}
#endif