spot/spot/twaalgos/powerset.hh
Alexandre Duret-Lutz 48edfd80c2 powerset: deal with accepting sinks more effectively
Part of #444.

* spot/twaalgos/powerset.cc, spot/twaalgos/powerset.hh: Implement
accepting sink handling.
* spot/twaalgos/minimize.cc (minimize_wdba): Pass sinks to
tgba_powerset.
* spot/misc/bitvect.hh: Add an interesects method.
* tests/core/ltl2tgba2.test: More tests.
* NEWS: Mention this new feature.
2020-12-08 17:55:50 +01:00

173 lines
5.8 KiB
C++

// -*- coding: utf-8 -*-
// Copyright (C) 2011, 2013-2015, 2019 Laboratoire de Recherche et
// Développement de l'Epita.
// Copyright (C) 2004 Laboratoire d'Informatique de Paris 6 (LIP6),
// département Systèmes Répartis Coopératifs (SRC), Université Pierre
// et Marie Curie.
//
// This file is part of Spot, a model checking library.
//
// Spot is free software; you can redistribute it and/or modify it
// under the terms of the GNU General Public License as published by
// the Free Software Foundation; either version 3 of the License, or
// (at your option) any later version.
//
// Spot is distributed in the hope that it will be useful, but WITHOUT
// ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
// or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public
// License for more details.
//
// You should have received a copy of the GNU General Public License
// along with this program. If not, see <http://www.gnu.org/licenses/>.
#pragma once
#include <set>
#include <vector>
#include <iosfwd>
#include <spot/twa/twagraph.hh>
namespace spot
{
struct SPOT_API power_map
{
typedef std::set<unsigned> power_state;
std::vector<power_state> map_;
const power_state&
states_of(unsigned s) const
{
return map_.at(s);
}
};
/// \brief Helper object to specify when an algorithm
/// should abort its construction.
class SPOT_API output_aborter
{
unsigned max_states_;
unsigned max_edges_;
mutable bool reason_is_states_;
public:
output_aborter(unsigned max_states,
unsigned max_edges = ~0U)
: max_states_(max_states), max_edges_(max_edges)
{
}
unsigned max_states() const
{
return max_states_;
}
unsigned max_edges() const
{
return max_edges_;
}
bool too_large(const const_twa_graph_ptr& aut) const
{
bool too_many_states = aut->num_states() > max_states_;
if (!too_many_states && (aut->num_edges() <= max_edges_))
return false;
// Only update the reason if we return true;
reason_is_states_ = too_many_states;
return true;
}
std::ostream& print_reason(std::ostream&) const;
};
/// \ingroup twa_misc
/// \brief Build a deterministic automaton, ignoring acceptance conditions.
///
/// This create a deterministic automaton that recognizes the
/// same language as \a aut would if its acceptance conditions
/// were ignored. This is the classical powerset algorithm.
///
/// If \a pm is supplied it will be filled with the set of original states
/// associated to each state of the deterministic automaton.
/// The \a merge argument can be set to false to prevent merging of
/// transitions.
///
/// If ab \a aborter is given, abort the construction whenever it
/// would build an automaton that is too large, and return nullptr.
///
/// If a vector of accepting sinks is given, all power-state that
/// contains any accepting sink will be merged into a single state
/// with number 0.
///
//@{
SPOT_API twa_graph_ptr
tgba_powerset(const const_twa_graph_ptr& aut,
power_map& pm, bool merge = true,
const output_aborter* aborter = nullptr,
std::vector<unsigned>* accepting_sinks = nullptr);
SPOT_API twa_graph_ptr
tgba_powerset(const const_twa_graph_ptr& aut,
const output_aborter* aborter = nullptr,
std::vector<unsigned>* accepting_sinks = nullptr);
//@}
/// \brief Determinize a TBA using the powerset construction.
///
/// The input automaton should have at most one acceptance
/// condition. Beware that not all Büchi automata can be
/// determinized, and this procedure does not ensure that the
/// produced automaton is equivalent to \a aut.
///
/// The construction is adapted from Section 3.2 of
/// \cite dax.07.atva
/// only adapted to work on TBA rather than BA.
///
/// If \a threshold_states is non null, abort the construction
/// whenever it would build an automaton that is more than \a
/// threshold_states time bigger (in term of states) than the
/// original automaton.
///
/// If \a threshold_cycles is non null, abort the construction
/// whenever an SCC of the constructed automaton has more than \a
/// threshold_cycles cycles.
SPOT_API twa_graph_ptr
tba_determinize(const const_twa_graph_ptr& aut,
unsigned threshold_states = 0,
unsigned threshold_cycles = 0);
/// \brief Determinize a TBA and make sure it is correct.
///
/// Apply tba_determinize(), then check that the result is
/// equivalent. If it isn't, return the original automaton.
///
/// Only one of \a f or \a neg_aut needs to be supplied. If
/// \a neg_aut is not given, it will be built from \a f.
///
/// \param aut the automaton to minimize
///
/// \param threshold_states if non null, abort the construction
/// whenever it would build an automaton that is more than \a
/// threshold time bigger (in term of states) than the original
/// automaton.
///
/// \param threshold_cycles can be used to abort the construction
/// if the number of cycles in a SCC of the constructed automaton
/// is bigger than the supplied value.
///
/// \param f the formula represented by the original automaton
///
/// \param neg_aut an automaton representing the negation of \a aut
///
/// \return a new tgba if the automaton could be determinized, \a aut if
/// the automaton cannot be determinized, 0 if we do not know if the
/// determinization is correct because neither \a f nor \a neg_aut
/// were supplied.
SPOT_API twa_graph_ptr
tba_determinize_check(const twa_graph_ptr& aut,
unsigned threshold_states = 0,
unsigned threshold_cycles = 0,
formula f = nullptr,
const_twa_graph_ptr neg_aut = nullptr);
}