mirrors/z3
3
0
Fork 0
mirror of https://github.com/Z3Prover/z3 synced 2025-04-27 02:45:51 +00:00
Code Activity

Merge branch 'master' of https://github.com/Z3Prover/z3

Browse source
This commit is contained in:
Christoph M. Wintersteiger 2017-05-12 14:29:23 +01:00
commit 45fc5d9fbd
60 changed files with 1217 additions and 1138 deletions
Download patch file Download diff file Expand all files Collapse all files

2
README-CMake.md
View file

@ -1,4 +1,4 @@
# Z3's CMake build system
# Z3's CMake build system
[CMake](https://cmake.org/) is a "meta build system" that reads a description
of the project written in the ``CMakeLists.txt`` files and emits a build

1
src/ast/rewriter/seq_rewriter.cpp
View file

@ -1506,6 +1506,7 @@ br_status seq_rewriter::mk_re_opt(expr* a, expr_ref& result) {
}
br_status seq_rewriter::mk_eq_core(expr * l, expr * r, expr_ref & result) {
TRACE("seq", tout << mk_pp(l, m()) << " = " << mk_pp(r, m()) << "\n";);
expr_ref_vector lhs(m()), rhs(m()), res(m());
bool changed = false;
if (!reduce_eq(l, r, lhs, rhs, changed)) {

30
src/ast/seq_decl_plugin.cpp
View file

@ -876,6 +876,36 @@ bool seq_decl_plugin::is_value(app* e) const {
}
}
bool seq_decl_plugin::are_equal(app* a, app* b) const {
if (a == b) return true;
// handle concatenations
return false;
}
bool seq_decl_plugin::are_distinct(app* a, app* b) const {
if (a == b) {
return false;
}
if (is_app_of(a, m_family_id, OP_STRING_CONST) &&
is_app_of(b, m_family_id, OP_STRING_CONST)) {
return true;
}
if (is_app_of(a, m_family_id, OP_SEQ_UNIT) &&
is_app_of(b, m_family_id, OP_SEQ_UNIT)) {
return true;
}
if (is_app_of(a, m_family_id, OP_SEQ_EMPTY) &&
is_app_of(b, m_family_id, OP_SEQ_UNIT)) {
return true;
}
if (is_app_of(b, m_family_id, OP_SEQ_EMPTY) &&
is_app_of(a, m_family_id, OP_SEQ_UNIT)) {
return true;
}
return false;
}
expr* seq_decl_plugin::get_some_value(sort* s) {
seq_util util(*m_manager);
if (util.is_seq(s)) {

6
src/ast/seq_decl_plugin.h
View file

@ -182,7 +182,11 @@ public:
virtual bool is_value(app * e) const;
virtual bool is_unique_value(app * e) const { return is_value(e); }
virtual bool is_unique_value(app * e) const { return false; }
virtual bool are_equal(app* a, app* b) const;
virtual bool are_distinct(app* a, app* b) const;
virtual expr * get_some_value(sort * s);

1
src/shell/lp_frontend.cpp
View file

@ -80,7 +80,6 @@ void run_solver(lp_params & params, char const * mps_file_name) {
solver->settings().set_message_ostream(&std::cout);
solver->settings().report_frequency = params.rep_freq();
solver->settings().print_statistics = params.print_stats();
solver->settings().presolve_with_double_solver_for_lar = params.presolve_with_dbl();
solver->find_maximal_solution();
*(solver->settings().get_message_ostream()) << "status is " << lp_status_to_string(solver->get_status()) << std::endl;

17
src/smt/smt_setup.cpp
View file

@ -471,12 +471,16 @@ namespace smt {
setup_r_arith();
}
void setup::setup_QF_LIRA(static_features const& st) {
setup_mi_arith();
}
void setup::setup_QF_LIA() {
TRACE("setup", tout << "setup_QF_LIA(st)\n";);
m_params.m_relevancy_lvl = 0;
m_params.m_arith_expand_eqs = true;
m_params.m_arith_reflect = false;
m_params.m_arith_propagate_eqs = false;
m_params.m_arith_propagate_eqs = false;
m_params.m_nnf_cnf = false;
setup_i_arith();
}
@ -720,10 +724,9 @@ namespace smt {
}
void setup::setup_r_arith() {
m_context.register_plugin(alloc(smt::theory_mi_arith, m_manager, m_params));
// Disabled in initial commit of LRA additions
// m_context.register_plugin(alloc(smt::theory_lra, m_manager, m_params));
// to disable theory lra
// m_context.register_plugin(alloc(smt::theory_mi_arith, m_manager, m_params));
m_context.register_plugin(alloc(smt::theory_lra, m_manager, m_params));
}
void setup::setup_mi_arith() {
@ -937,7 +940,9 @@ namespace smt {
}
if (st.num_theories() == 1 && is_arith(st)) {
if (st.m_has_real)
if ((st.m_has_int && st.m_has_real) || (st.m_num_non_linear != 0))
setup_QF_LIRA(st);
else if (st.m_has_real)
setup_QF_LRA(st);
else
setup_QF_LIA(st);

1
src/smt/smt_setup.h
View file

@ -65,6 +65,7 @@ namespace smt {
void setup_QF_LRA();
void setup_QF_LRA(static_features const & st);
void setup_QF_LIA();
void setup_QF_LIRA(static_features const& st);
void setup_QF_LIA(static_features const & st);
void setup_QF_UFLIA();
void setup_QF_UFLIA(static_features & st);

11
src/smt/tactic/smt_tactic.cpp
View file

@ -21,6 +21,7 @@ Notes:
#include"smt_kernel.h"
#include"smt_params.h"
#include"smt_params_helper.hpp"
#include"lp_params.hpp"
#include"rewriter_types.h"
#include"filter_model_converter.h"
#include"ast_util.h"
@ -64,6 +65,10 @@ public:
return m_params;
}
params_ref & params() {
return m_params_ref;
}
void updt_params_core(params_ref const & p) {
m_candidate_models = p.get_bool("candidate_models", false);
m_fail_if_inconclusive = p.get_bool("fail_if_inconclusive", true);
@ -73,6 +78,7 @@ public:
TRACE("smt_tactic", tout << "updt_params: " << p << "\n";);
updt_params_core(p);
fparams().updt_params(p);
m_params_ref.copy(p);
m_logic = p.get_sym(symbol("logic"), m_logic);
if (m_logic != symbol::null && m_ctx) {
m_ctx->set_logic(m_logic);
@ -84,6 +90,7 @@ public:
r.insert("candidate_models", CPK_BOOL, "(default: false) create candidate models even when quantifier or theory reasoning is incomplete.");
r.insert("fail_if_inconclusive", CPK_BOOL, "(default: true) fail if found unsat (sat) for under (over) approximated goal.");
smt_params_helper::collect_param_descrs(r);
lp_params::collect_param_descrs(r);
}
@ -112,10 +119,12 @@ public:
struct scoped_init_ctx {
smt_tactic & m_owner;
smt_params m_params; // smt-setup overwrites parameters depending on the current assertions.
params_ref m_params_ref;
scoped_init_ctx(smt_tactic & o, ast_manager & m):m_owner(o) {
m_params = o.fparams();
smt::kernel * new_ctx = alloc(smt::kernel, m, m_params);
m_params_ref = o.params();
smt::kernel * new_ctx = alloc(smt::kernel, m, m_params, m_params_ref);
TRACE("smt_tactic", tout << "logic: " << o.m_logic << "\n";);
new_ctx->set_logic(o.m_logic);
if (o.m_callback) {

57
src/smt/theory_lra.cpp
View file

@ -143,14 +143,11 @@ namespace smt {
theory_lra& th;
ast_manager& m;
theory_arith_params& m_arith_params;
lp_params m_lp_params; // seeded from global parameters.
arith_util a;
arith_eq_adapter m_arith_eq_adapter;
vector<rational> m_columns;
int m_print_counter = 0;
// temporary values kept during internalization
struct internalize_state {
expr_ref_vector m_terms;
@ -282,14 +279,15 @@ namespace smt {
expr* get_owner(theory_var v) const { return get_enode(v)->get_owner(); }
void init_solver() {
if (m_solver) return;
lp_params lp(ctx().get_params());
m_solver = alloc(lean::lar_solver);
m_theory_var2var_index.reset();
m_solver->settings().set_resource_limit(m_resource_limit);
m_solver->settings().simplex_strategy() = static_cast<lean::simplex_strategy_enum>(m_lp_params.simplex_strategy());
m_solver->settings().presolve_with_double_solver_for_lar = m_lp_params.presolve_with_dbl();
m_solver->settings().simplex_strategy() = static_cast<lean::simplex_strategy_enum>(lp.simplex_strategy());
reset_variable_values();
m_solver->settings().bound_propagation() = BP_NONE != propagation_mode();
m_solver->set_propagate_bounds_on_pivoted_rows_mode(m_lp_params.bprop_on_pivoted_rows());
m_solver->set_propagate_bounds_on_pivoted_rows_mode(lp.bprop_on_pivoted_rows());
//m_solver->settings().set_ostream(0);
}
@ -646,9 +644,11 @@ namespace smt {
public:
imp(theory_lra& th, ast_manager& m, theory_arith_params& p):
th(th), m(m), m_arith_params(p), a(m),
m_arith_eq_adapter(th, p, a),
imp(theory_lra& th, ast_manager& m, theory_arith_params& ap):
th(th), m(m),
m_arith_params(ap),
a(m),
m_arith_eq_adapter(th, ap, a),
m_internalize_head(0),
m_delay_constraints(false),
m_delayed_terms(m),
@ -657,14 +657,18 @@ namespace smt {
m_assume_eq_head(0),
m_num_conflicts(0),
m_model_eqs(DEFAULT_HASHTABLE_INITIAL_CAPACITY, var_value_hash(*this), var_value_eq(*this)),
m_solver(0),
m_resource_limit(*this) {
init_solver();
}
~imp() {
del_bounds(0);
std::for_each(m_internalize_states.begin(), m_internalize_states.end(), delete_proc<internalize_state>());
}
void init(context* ctx) {
init_solver();
}
bool internalize_atom(app * atom, bool gate_ctx) {
if (m_delay_constraints) {
@ -1978,7 +1982,6 @@ namespace smt {
typedef pair_hash<obj_hash<rational>, bool_hash> value_sort_pair_hash;
typedef map<value_sort_pair, theory_var, value_sort_pair_hash, default_eq<value_sort_pair> > value2var;
value2var m_fixed_var_table;
const lean::constraint_index null_index = static_cast<lean::constraint_index>(-1);
void propagate_eqs(lean::var_index vi, lean::constraint_index ci, lean::lconstraint_kind k, lp::bound& b) {
if (propagate_eqs()) {
@ -2008,7 +2011,7 @@ namespace smt {
bool proofs_enabled() const { return m.proofs_enabled(); }
bool use_tableau() const { return m_lp_params.simplex_strategy() < 2; }
bool use_tableau() const { return lp_params(ctx().get_params()).simplex_strategy() < 2; }
void set_upper_bound(lean::var_index vi, lean::constraint_index ci, rational const& v) { set_bound(vi, ci, v, false); }
@ -2022,10 +2025,10 @@ namespace smt {
lean::var_index ti = m_solver->adjust_term_index(vi);
auto& vec = is_lower ? m_lower_terms : m_upper_terms;
if (vec.size() <= ti) {
vec.resize(ti + 1, constraint_bound(null_index, rational()));
vec.resize(ti + 1, constraint_bound(UINT_MAX, rational()));
}
constraint_bound& b = vec[ti];
if (b.first == null_index || (is_lower? b.second < v : b.second > v)) {
if (b.first == UINT_MAX || (is_lower? b.second < v : b.second > v)) {
ctx().push_trail(vector_value_trail<context, constraint_bound>(vec, ti));
b.first = ci;
b.second = v;
@ -2045,7 +2048,7 @@ namespace smt {
rational val;
TRACE("arith", tout << vi << " " << v << "\n";);
if (v != null_theory_var && a.is_numeral(get_owner(v), val) && bound == val) {
ci = null_constraint_index;
ci = UINT_MAX;
return bound == val;
}
@ -2053,7 +2056,7 @@ namespace smt {
if (vec.size() > ti) {
constraint_bound& b = vec[ti];
ci = b.first;
return ci != null_index && bound == b.second;
return ci != UINT_MAX && bound == b.second;
}
else {
return false;
@ -2137,22 +2140,10 @@ namespace smt {
if (m_solver->A_r().row_count() > m_stats.m_max_rows)
m_stats.m_max_rows = m_solver->A_r().row_count();
TRACE("arith_verbose", display(tout););
bool print = false && m_print_counter++ % 1000 == 0;
stopwatch sw;
if (print) {
sw.start();
}
lean::lp_status status = m_solver->find_feasible_solution();
if (print) {
sw.stop();
}
m_stats.m_num_iterations = m_solver->settings().st().m_total_iterations;
m_stats.m_num_factorizations = m_solver->settings().st().m_num_factorizations;
m_stats.m_need_to_solve_inf = m_solver->settings().st().m_need_to_solve_inf;
if (print) {
IF_VERBOSE(0, verbose_stream() << status << " " << sw.get_seconds() << " " << m_stats.m_num_iterations << " " << m_print_counter << "\n";);
}
//m_stats.m_num_iterations_with_no_progress += m_solver->settings().st().m_iters_with_no_cost_growing;
switch (status) {
case lean::lp_status::INFEASIBLE:
@ -2175,10 +2166,11 @@ namespace smt {
literal_vector m_core;
svector<enode_pair> m_eqs;
vector<parameter> m_params;
lean::constraint_index const null_constraint_index = UINT_MAX;
// lean::constraint_index const null_constraint_index = UINT_MAX; // not sure what a correct fix is
void set_evidence(lean::constraint_index idx) {
if (idx == null_constraint_index) {
if (idx == UINT_MAX) {
return;
}
switch (m_constraint_sources[idx]) {
@ -2517,9 +2509,9 @@ namespace smt {
}
};
theory_lra::theory_lra(ast_manager& m, theory_arith_params& p):
theory_lra::theory_lra(ast_manager& m, theory_arith_params& ap):
theory(m.get_family_id("arith")) {
m_imp = alloc(imp, *this, m, p);
m_imp = alloc(imp, *this, m, ap);
}
theory_lra::~theory_lra() {
dealloc(m_imp);
@ -2529,6 +2521,7 @@ namespace smt {
}
void theory_lra::init(context * ctx) {
theory::init(ctx);
m_imp->init(ctx);
}
bool theory_lra::internalize_atom(app * atom, bool gate_ctx) {
return m_imp->internalize_atom(atom, gate_ctx);

3
src/smt/theory_lra.h
View file

@ -26,8 +26,9 @@ namespace smt {
class theory_lra : public theory, public theory_opt {
class imp;
imp* m_imp;
public:
theory_lra(ast_manager& m, theory_arith_params& params);
theory_lra(ast_manager& m, theory_arith_params& ap);
virtual ~theory_lra();
virtual theory* mk_fresh(context* new_ctx);
virtual char const* get_name() const { return "lra"; }

14
src/tactic/smtlogics/qflra_tactic.cpp
View file

@ -22,7 +22,6 @@ Notes:
#include"solve_eqs_tactic.h"
#include"elim_uncnstr_tactic.h"
#include"smt_tactic.h"
// include"mip_tactic.h"
#include"recover_01_tactic.h"
#include"ctx_simplify_tactic.h"
#include"probe_arith.h"
@ -72,5 +71,18 @@ tactic * mk_qflra_tactic(ast_manager & m, params_ref const & p) {
// using_params(mk_smt_tactic(), pivot_p)),
// p);
#if 0
params_ref simplex_0, simplex_1, simplex_2;
simplex_0.set_uint("lp.simplex_strategy", 0);
simplex_1.set_uint("lp.simplex_strategy", 1);
simplex_2.set_uint("lp.simplex_strategy", 2);
return par(using_params(mk_smt_tactic(), simplex_0),
using_params(mk_smt_tactic(), simplex_1),
using_params(mk_smt_tactic(), simplex_2));
#else
return using_params(using_params(mk_smt_tactic(), pivot_p), p);
#endif
}

8
src/test/lp.cpp
View file

@ -34,6 +34,10 @@ Author: Lev Nachmanson
namespace lean {
unsigned seed = 1;
random_gen g_rand;
static unsigned my_random() {
return g_rand();
}
struct simple_column_namer:public column_namer
{
std::string get_column_name(unsigned j) const override {
@ -1077,7 +1081,7 @@ bool get_double_from_args_parser(const char * option, argument_parser & args_par
void update_settings(argument_parser & args_parser, lp_settings& settings) {
unsigned n;
settings.m_simplex_strategy = simplex_strategy_enum::no_tableau;
settings.m_simplex_strategy = simplex_strategy_enum::lu;
if (get_int_from_args_parser("--rep_frq", args_parser, n))
settings.report_frequency = n;
else
@ -1104,7 +1108,7 @@ void update_settings(argument_parser & args_parser, lp_settings& settings) {
settings.harris_feasibility_tolerance = d;
}
if (get_int_from_args_parser("--random_seed", args_parser, n)) {
settings.random_seed = n;
settings.random_seed(n);
}
if (get_int_from_args_parser("--simplex_strategy", args_parser, n)) {
settings.simplex_strategy() = static_cast<simplex_strategy_enum>(n);

24
src/test/smt_reader.h
View file

@ -56,23 +56,27 @@ namespace lean {
struct formula_constraint {
lconstraint_kind m_kind;
std::vector<std::pair<mpq, std::string>> m_coeffs;
mpq m_right_side = numeric_traits<mpq>::zero();
mpq m_right_side;
void add_pair(mpq c, std::string name) {
m_coeffs.push_back(make_pair(c, name));
}
formula_constraint() : m_right_side(numeric_traits<mpq>::zero()) {}
};
lisp_elem m_formula_lisp_elem;
std::unordered_map<std::string, unsigned> m_name_to_var_index;
std::vector<formula_constraint> m_constraints;
std::string m_file_name;
std::ifstream m_file_stream;
std::string m_line;
bool m_is_OK = true;
unsigned m_line_number = 0;
smt_reader(std::string file_name):
m_file_name(file_name), m_file_stream(file_name) {
std::vector<formula_constraint> m_constraints;
bool m_is_OK;
unsigned m_line_number;
std::string m_file_name;
std::ifstream m_file_stream;
std::string m_line;
smt_reader(std::string file_name):
m_is_OK(true),
m_line_number(0),
m_file_name(file_name),
m_file_stream(file_name) {
}
void set_error() {
@ -364,7 +368,7 @@ namespace lean {
if (it!= m_name_to_var_index.end())
return it->second;
unsigned ret= m_name_to_var_index.size();
unsigned ret = static_cast<unsigned>(m_name_to_var_index.size());
m_name_to_var_index[s] = ret;
return ret;
}

2
src/util/hash.cpp
View file

@ -83,8 +83,8 @@ unsigned string_hash(const char * str, unsigned length, unsigned init_value) {
Z3_fallthrough;
case 1 :
a+=str[0];
Z3_fallthrough;
/* case 0: nothing left to add */
break;
}
mix(a,b,c);
/*-------------------------------------------- report the result */

5
src/util/lp/binary_heap_priority_queue.h
View file

@ -16,8 +16,7 @@ class binary_heap_priority_queue {
// indexing for A starts from 1
vector<unsigned> m_heap; // keeps the elements of the queue
vector<int> m_heap_inverse; // o = m_heap[m_heap_inverse[o]]
unsigned m_heap_size = 0;
unsigned m_heap_size;
// is is the child place in heap
void swap_with_parent(unsigned i);
void put_at(unsigned i, unsigned h);
@ -29,7 +28,7 @@ public:
public:
void remove(unsigned o);
unsigned size() const { return m_heap_size; }
binary_heap_priority_queue(): m_heap(1) {} // the empty constructror
binary_heap_priority_queue(): m_heap(1), m_heap_size(0) {} // the empty constructror
// n is the initial queue capacity.
// The capacity will be enlarged two times automatically if needed
binary_heap_priority_queue(unsigned n);

3
src/util/lp/binary_heap_priority_queue.hpp
View file

@ -83,7 +83,8 @@ template <typename T> void binary_heap_priority_queue<T>::remove(unsigned o) {
template <typename T> binary_heap_priority_queue<T>::binary_heap_priority_queue(unsigned n) :
m_priorities(n),
m_heap(n + 1), // because the indexing for A starts from 1
m_heap_inverse(n, -1)
m_heap_inverse(n, -1),
m_heap_size(0)
{ }

4
src/util/lp/binary_heap_upair_queue.h
View file

@ -20,8 +20,8 @@ template <typename T>
class binary_heap_upair_queue {
binary_heap_priority_queue<T> m_q;
std::unordered_map<upair, unsigned> m_pairs_to_index;
vector<upair> m_pairs; // inverse to index
vector<unsigned> m_available_spots;
svector<upair> m_pairs; // inverse to index
svector<unsigned> m_available_spots;
public:
binary_heap_upair_queue(unsigned size);

20
src/util/lp/bound_analyzer_on_row.h
View file

@ -18,12 +18,13 @@ namespace lean {
class bound_analyzer_on_row {
linear_combination_iterator<mpq> & m_it;
unsigned m_row_or_term_index;
int m_column_of_u = -1; // index of an unlimited from above monoid
// -1 means that such a value is not found, -2 means that at least two of such monoids were found
int m_column_of_l = -1; // index of an unlimited from below monoid
impq m_rs;
bound_propagator & m_bp;
bound_propagator & m_bp;
unsigned m_row_or_term_index;
int m_column_of_u; // index of an unlimited from above monoid
// -1 means that such a value is not found, -2 means that at least two of such monoids were found
int m_column_of_l; // index of an unlimited from below monoid
impq m_rs;
public :
// constructor
bound_analyzer_on_row(
@ -34,9 +35,11 @@ public :
)
:
m_it(it),
m_bp(bp),
m_row_or_term_index(row_or_term_index),
m_rs(rs),
m_bp(bp)
m_column_of_u(-1),
m_column_of_l(-1),
m_rs(rs)
{}
@ -250,7 +253,6 @@ public :
if (str)
strict = true;
}
bound /= l_coeff;
if (is_pos(l_coeff)) {
limit_j(m_column_of_l, bound, true, false, strict);

45
src/util/lp/column_info.h
View file

@ -15,26 +15,26 @@ inline bool is_valid(unsigned j) { return static_cast<int>(j) >= 0;}
template <typename T>
class column_info {
std::string m_name;
bool m_low_bound_is_set = false;
bool m_low_bound_is_strict = false;
bool m_upper_bound_is_set = false;
bool m_upper_bound_is_strict = false;
T m_low_bound;
T m_upper_bound;
T m_cost = numeric_traits<T>::zero();
T m_fixed_value;
bool m_is_fixed = false;
unsigned m_column_index = static_cast<unsigned>(-1);
bool m_low_bound_is_set;
bool m_low_bound_is_strict;
bool m_upper_bound_is_set;
bool m_upper_bound_is_strict;
T m_low_bound;
T m_upper_bound;
T m_fixed_value;
bool m_is_fixed;
T m_cost;
unsigned m_column_index;
public:
bool operator==(const column_info & c) const {
return m_name == c.m_name &&
return m_name == c.m_name &&
m_low_bound_is_set == c.m_low_bound_is_set &&
m_low_bound_is_strict == c.m_low_bound_is_strict &&
m_upper_bound_is_set == c.m_upper_bound_is_set&&
m_upper_bound_is_strict == c.m_upper_bound_is_strict&&
(!m_low_bound_is_set || m_low_bound == c.m_low_bound) &&
(!m_upper_bound_is_set || m_upper_bound == c.m_upper_bound) &&
m_cost == c.m_cost&&
m_cost == c.m_cost &&
m_is_fixed == c.m_is_fixed &&
(!m_is_fixed || m_fixed_value == c.m_fixed_value) &&
m_column_index == c.m_column_index;
@ -44,9 +44,24 @@ public:
m_column_index = j;
}
// the default constructor
column_info() {}
column_info(unsigned column_index) : m_column_index(column_index) {
column_info():
m_low_bound_is_set(false),
m_low_bound_is_strict(false),
m_upper_bound_is_set (false),
m_upper_bound_is_strict (false),
m_is_fixed(false),
m_cost(numeric_traits<T>::zero()),
m_column_index(static_cast<unsigned>(-1))
{}
column_info(unsigned column_index) :
m_low_bound_is_set(false),
m_low_bound_is_strict(false),
m_upper_bound_is_set (false),
m_upper_bound_is_strict (false),
m_is_fixed(false),
m_cost(numeric_traits<T>::zero()),
m_column_index(column_index) {
}
column_info(const column_info & ci) {

43
src/util/lp/conversion_helper.h Normal file
View file

@ -0,0 +1,43 @@
/*
Copyright (c) 2013 Microsoft Corporation. All rights reserved.
Author: Lev Nachmanson
*/
#pragma once
namespace lean {
template <typename V>
struct conversion_helper {
static V get_low_bound(const column_info<mpq> & ci) {
return V(ci.get_low_bound(), ci.low_bound_is_strict()? 1 : 0);
}
static V get_upper_bound(const column_info<mpq> & ci) {
return V(ci.get_upper_bound(), ci.upper_bound_is_strict()? -1 : 0);
}
};
template<>
struct conversion_helper <double> {
static double get_upper_bound(const column_info<mpq> & ci) {
if (!ci.upper_bound_is_strict())
return ci.get_upper_bound().get_double();
double eps = 0.00001;
if (!ci.low_bound_is_set())
return ci.get_upper_bound().get_double() - eps;
eps = std::min((ci.get_upper_bound() - ci.get_low_bound()).get_double() / 1000, eps);
return ci.get_upper_bound().get_double() - eps;
}
static double get_low_bound(const column_info<mpq> & ci) {
if (!ci.low_bound_is_strict())
return ci.get_low_bound().get_double();
double eps = 0.00001;
if (!ci.upper_bound_is_set())
return ci.get_low_bound().get_double() + eps;
eps = std::min((ci.get_upper_bound() - ci.get_low_bound()).get_double() / 1000, eps);
return ci.get_low_bound().get_double() + eps;
}
};
}

11
src/util/lp/core_solver_pretty_printer.h
View file

@ -16,7 +16,6 @@ template <typename T, typename X> class lp_core_solver_base; // forward definiti
template <typename T, typename X>
class core_solver_pretty_printer {
std::ostream & m_out;
template<typename A> using vector = vector<A>;
typedef std::string string;
lp_core_solver_base<T, X> & m_core_solver;
vector<unsigned> m_column_widths;
@ -34,15 +33,15 @@ class core_solver_pretty_printer {
std::string m_cost_title;
std::string m_basis_heading_title;
std::string m_x_title;
std::string m_low_bounds_title = "low";
std::string m_upp_bounds_title = "upp";
std::string m_exact_norm_title = "exact cn";
std::string m_approx_norm_title = "approx cn";
std::string m_low_bounds_title;
std::string m_upp_bounds_title;
std::string m_exact_norm_title;
std::string m_approx_norm_title;
unsigned ncols() { return m_core_solver.m_A.column_count(); }
unsigned nrows() { return m_core_solver.m_A.row_count(); }
unsigned m_artificial_start = std::numeric_limits<unsigned>::max();
unsigned m_artificial_start;
indexed_vector<T> m_w_buff;
indexed_vector<T> m_ed_buff;
vector<T> m_exact_column_norms;

6
src/util/lp/core_solver_pretty_printer.hpp
View file

@ -23,6 +23,12 @@ core_solver_pretty_printer<T, X>::core_solver_pretty_printer(lp_core_solver_base
m_rs(ncols(), zero_of_type<X>()),
m_w_buff(core_solver.m_w),
m_ed_buff(core_solver.m_ed) {
m_low_bounds_title = "low";
m_upp_bounds_title = "upp";
m_exact_norm_title = "exact cn";
m_approx_norm_title = "approx cn";
m_artificial_start = std::numeric_limits<unsigned>::max();
m_column_widths.resize(core_solver.m_A.column_count(), 0),
init_m_A_and_signs();
init_costs();

576
src/util/lp/init_lar_solver.h Normal file
View file

@ -0,0 +1,576 @@
/*
Copyright (c) 2017 Microsoft Corporation
Author: Lev Nachmanson
*/
// here we are inside lean::lar_solver class
bool strategy_is_undecided() const {
return m_settings.simplex_strategy() == simplex_strategy_enum::undecided;
}
var_index add_var(unsigned ext_j) {
var_index i;
lean_assert (ext_j < m_terms_start_index);
if (ext_j >= m_terms_start_index)
throw 0; // todo : what is the right way to exit?
if (try_get_val(m_ext_vars_to_columns, ext_j, i)) {
return i;
}
lean_assert(m_vars_to_ul_pairs.size() == A_r().column_count());
i = A_r().column_count();
m_vars_to_ul_pairs.push_back (ul_pair(static_cast<unsigned>(-1)));
add_non_basic_var_to_core_fields(ext_j);
lean_assert(sizes_are_correct());
return i;
}
void register_new_ext_var_index(unsigned ext_v) {
lean_assert(!contains(m_ext_vars_to_columns, ext_v));
unsigned j = static_cast<unsigned>(m_ext_vars_to_columns.size());
m_ext_vars_to_columns[ext_v] = j;
lean_assert(m_columns_to_ext_vars_or_term_indices.size() == j);
m_columns_to_ext_vars_or_term_indices.push_back(ext_v);
}
void add_non_basic_var_to_core_fields(unsigned ext_j) {
register_new_ext_var_index(ext_j);
m_mpq_lar_core_solver.m_column_types.push_back(column_type::free_column);
m_columns_with_changed_bound.increase_size_by_one();
add_new_var_to_core_fields_for_mpq(false);
if (use_lu())
add_new_var_to_core_fields_for_doubles(false);
}
void add_new_var_to_core_fields_for_doubles(bool register_in_basis) {
unsigned j = A_d().column_count();
A_d().add_column();
lean_assert(m_mpq_lar_core_solver.m_d_x.size() == j);
// lean_assert(m_mpq_lar_core_solver.m_d_low_bounds.size() == j && m_mpq_lar_core_solver.m_d_upper_bounds.size() == j); // restore later
m_mpq_lar_core_solver.m_d_x.resize(j + 1 );
m_mpq_lar_core_solver.m_d_low_bounds.resize(j + 1);
m_mpq_lar_core_solver.m_d_upper_bounds.resize(j + 1);
lean_assert(m_mpq_lar_core_solver.m_d_heading.size() == j); // as A().column_count() on the entry to the method
if (register_in_basis) {
A_d().add_row();
m_mpq_lar_core_solver.m_d_heading.push_back(m_mpq_lar_core_solver.m_d_basis.size());
m_mpq_lar_core_solver.m_d_basis.push_back(j);
}else {
m_mpq_lar_core_solver.m_d_heading.push_back(- static_cast<int>(m_mpq_lar_core_solver.m_d_nbasis.size()) - 1);
m_mpq_lar_core_solver.m_d_nbasis.push_back(j);
}
}
void add_new_var_to_core_fields_for_mpq(bool register_in_basis) {
unsigned j = A_r().column_count();
A_r().add_column();
lean_assert(m_mpq_lar_core_solver.m_r_x.size() == j);
// lean_assert(m_mpq_lar_core_solver.m_r_low_bounds.size() == j && m_mpq_lar_core_solver.m_r_upper_bounds.size() == j); // restore later
m_mpq_lar_core_solver.m_r_x.resize(j + 1);
m_mpq_lar_core_solver.m_r_low_bounds.increase_size_by_one();
m_mpq_lar_core_solver.m_r_upper_bounds.increase_size_by_one();
m_mpq_lar_core_solver.m_r_solver.m_inf_set.increase_size_by_one();
m_mpq_lar_core_solver.m_r_solver.m_costs.resize(j + 1);
m_mpq_lar_core_solver.m_r_solver.m_d.resize(j + 1);
lean_assert(m_mpq_lar_core_solver.m_r_heading.size() == j); // as A().column_count() on the entry to the method
if (register_in_basis) {
A_r().add_row();
m_mpq_lar_core_solver.m_r_heading.push_back(m_mpq_lar_core_solver.m_r_basis.size());
m_mpq_lar_core_solver.m_r_basis.push_back(j);
if (m_settings.bound_propagation())
m_rows_with_changed_bounds.insert(A_r().row_count() - 1);
} else {
m_mpq_lar_core_solver.m_r_heading.push_back(- static_cast<int>(m_mpq_lar_core_solver.m_r_nbasis.size()) - 1);
m_mpq_lar_core_solver.m_r_nbasis.push_back(j);
}
}
var_index add_term_undecided(const vector<std::pair<mpq, var_index>> & coeffs,
const mpq &m_v) {
m_terms.push_back(new lar_term(coeffs, m_v));
m_orig_terms.push_back(new lar_term(coeffs, m_v));
return m_terms_start_index + m_terms.size() - 1;
}
// terms
var_index add_term(const vector<std::pair<mpq, var_index>> & coeffs,
const mpq &m_v) {
if (strategy_is_undecided())
return add_term_undecided(coeffs, m_v);
m_terms.push_back(new lar_term(coeffs, m_v));
m_orig_terms.push_back(new lar_term(coeffs, m_v));
unsigned adjusted_term_index = m_terms.size() - 1;
var_index ret = m_terms_start_index + adjusted_term_index;
if (use_tableau() && !coeffs.empty()) {
add_row_for_term(m_orig_terms.back(), ret);
if (m_settings.bound_propagation())
m_rows_with_changed_bounds.insert(A_r().row_count() - 1);
}
lean_assert(m_ext_vars_to_columns.size() == A_r().column_count());
return ret;
}
void add_row_for_term(const lar_term * term, unsigned term_ext_index) {
lean_assert(sizes_are_correct());
add_row_from_term_no_constraint(term, term_ext_index);
lean_assert(sizes_are_correct());
}
void add_row_from_term_no_constraint(const lar_term * term, unsigned term_ext_index) {
register_new_ext_var_index(term_ext_index);
// j will be a new variable
unsigned j = A_r().column_count();
ul_pair ul(j);
m_vars_to_ul_pairs.push_back(ul);
add_basic_var_to_core_fields();
if (use_tableau()) {
auto it = iterator_on_term_with_basis_var(*term, j);
A_r().fill_last_row_with_pivoting(it,
m_mpq_lar_core_solver.m_r_solver.m_basis_heading);
m_mpq_lar_core_solver.m_r_solver.m_b.resize(A_r().column_count(), zero_of_type<mpq>());
} else {
fill_last_row_of_A_r(A_r(), term);
}
m_mpq_lar_core_solver.m_r_x[j] = get_basic_var_value_from_row_directly(A_r().row_count() - 1);
if (use_lu())
fill_last_row_of_A_d(A_d(), term);
}
void add_basic_var_to_core_fields() {
bool use_lu = m_mpq_lar_core_solver.need_to_presolve_with_double_solver();
lean_assert(!use_lu || A_r().column_count() == A_d().column_count());
m_mpq_lar_core_solver.m_column_types.push_back(column_type::free_column);
m_columns_with_changed_bound.increase_size_by_one();
m_rows_with_changed_bounds.increase_size_by_one();
add_new_var_to_core_fields_for_mpq(true);
if (use_lu)
add_new_var_to_core_fields_for_doubles(true);
}
constraint_index add_var_bound(var_index j, lconstraint_kind kind, const mpq & right_side) {
constraint_index ci = m_constraints.size();
if (!is_term(j)) { // j is a var
auto vc = new lar_var_constraint(j, kind, right_side);
m_constraints.push_back(vc);
update_column_type_and_bound(j, kind, right_side, ci);
} else {
add_var_bound_on_constraint_for_term(j, kind, right_side, ci);
}
lean_assert(sizes_are_correct());
return ci;
}
void update_column_type_and_bound(var_index j, lconstraint_kind kind, const mpq & right_side, constraint_index constr_index) {
switch(m_mpq_lar_core_solver.m_column_types[j]) {
case column_type::free_column:
update_free_column_type_and_bound(j, kind, right_side, constr_index);
break;
case column_type::boxed:
update_boxed_column_type_and_bound(j, kind, right_side, constr_index);
break;
case column_type::low_bound:
update_low_bound_column_type_and_bound(j, kind, right_side, constr_index);
break;
case column_type::upper_bound:
update_upper_bound_column_type_and_bound(j, kind, right_side, constr_index);
break;
case column_type::fixed:
update_fixed_column_type_and_bound(j, kind, right_side, constr_index);
break;
default:
lean_assert(false); // cannot be here
}
}
void add_var_bound_on_constraint_for_term(var_index j, lconstraint_kind kind, const mpq & right_side, constraint_index ci) {
lean_assert(is_term(j));
unsigned adjusted_term_index = adjust_term_index(j);
unsigned term_j;
if (try_get_val(m_ext_vars_to_columns, j, term_j)) {
mpq rs = right_side - m_orig_terms[adjusted_term_index]->m_v;
m_constraints.push_back(new lar_term_constraint(m_orig_terms[adjusted_term_index], kind, right_side));
update_column_type_and_bound(term_j, kind, rs, ci);
}
else {
add_constraint_from_term_and_create_new_column_row(j, m_orig_terms[adjusted_term_index], kind, right_side);
}
}
void add_constraint_from_term_and_create_new_column_row(unsigned term_j, const lar_term* term,
lconstraint_kind kind, const mpq & right_side) {
add_row_from_term_no_constraint(term, term_j);
unsigned j = A_r().column_count() - 1;
update_column_type_and_bound(j, kind, right_side - term->m_v, m_constraints.size());
m_constraints.push_back(new lar_term_constraint(term, kind, right_side));
lean_assert(A_r().column_count() == m_mpq_lar_core_solver.m_r_solver.m_costs.size());
}
void decide_on_strategy_and_adjust_initial_state() {
lean_assert(strategy_is_undecided());
if (m_vars_to_ul_pairs.size() > m_settings.column_number_threshold_for_using_lu_in_lar_solver) {
m_settings.simplex_strategy() = simplex_strategy_enum::lu;
} else {
m_settings.simplex_strategy() = simplex_strategy_enum::tableau_rows; // todo: when to switch to tableau_costs?
}
adjust_initial_state();
}
void adjust_initial_state() {
switch (m_settings.simplex_strategy()) {
case simplex_strategy_enum::lu:
adjust_initial_state_for_lu();
break;
case simplex_strategy_enum::tableau_rows:
adjust_initial_state_for_tableau_rows();
break;
case simplex_strategy_enum::tableau_costs:
lean_assert(false); // not implemented
case simplex_strategy_enum::undecided:
adjust_initial_state_for_tableau_rows();
break;
}
}
void adjust_initial_state_for_lu() {
copy_from_mpq_matrix(A_d());
unsigned n = A_d().column_count();
m_mpq_lar_core_solver.m_d_x.resize(n);
m_mpq_lar_core_solver.m_d_low_bounds.resize(n);
m_mpq_lar_core_solver.m_d_upper_bounds.resize(n);
m_mpq_lar_core_solver.m_d_heading = m_mpq_lar_core_solver.m_r_heading;
m_mpq_lar_core_solver.m_d_basis = m_mpq_lar_core_solver.m_r_basis;
/*
unsigned j = A_d().column_count();
A_d().add_column();
lean_assert(m_mpq_lar_core_solver.m_d_x.size() == j);
// lean_assert(m_mpq_lar_core_solver.m_d_low_bounds.size() == j && m_mpq_lar_core_solver.m_d_upper_bounds.size() == j); // restore later
m_mpq_lar_core_solver.m_d_x.resize(j + 1 );
m_mpq_lar_core_solver.m_d_low_bounds.resize(j + 1);
m_mpq_lar_core_solver.m_d_upper_bounds.resize(j + 1);
lean_assert(m_mpq_lar_core_solver.m_d_heading.size() == j); // as A().column_count() on the entry to the method
if (register_in_basis) {
A_d().add_row();
m_mpq_lar_core_solver.m_d_heading.push_back(m_mpq_lar_core_solver.m_d_basis.size());
m_mpq_lar_core_solver.m_d_basis.push_back(j);
}else {
m_mpq_lar_core_solver.m_d_heading.push_back(- static_cast<int>(m_mpq_lar_core_solver.m_d_nbasis.size()) - 1);
m_mpq_lar_core_solver.m_d_nbasis.push_back(j);
}*/
}
void adjust_initial_state_for_tableau_rows() {
for (unsigned j = 0; j < m_terms.size(); j++) {
if (contains(m_ext_vars_to_columns, j + m_terms_start_index))
continue;
add_row_from_term_no_constraint(m_terms[j], j + m_terms_start_index);
}
}
// this fills the last row of A_d and sets the basis column: -1 in the last column of the row
void fill_last_row_of_A_d(static_matrix<double, double> & A, const lar_term* ls) {
lean_assert(A.row_count() > 0);
lean_assert(A.column_count() > 0);
unsigned last_row = A.row_count() - 1;
lean_assert(A.m_rows[last_row].empty());
for (auto & t : ls->m_coeffs) {
lean_assert(!is_zero(t.second));
var_index j = t.first;
A.set(last_row, j, - t.second.get_double());
}
unsigned basis_j = A.column_count() - 1;
A.set(last_row, basis_j, - 1 );
}
void update_free_column_type_and_bound(var_index j, lconstraint_kind kind, const mpq & right_side, constraint_index constr_ind) {
mpq y_of_bound(0);
switch (kind) {
case LT:
y_of_bound = -1;
case LE:
m_mpq_lar_core_solver.m_column_types[j] = column_type::upper_bound;
lean_assert(m_mpq_lar_core_solver.m_column_types()[j] == column_type::upper_bound);
lean_assert(m_mpq_lar_core_solver.m_r_upper_bounds.size() > j);
{
auto up = numeric_pair<mpq>(right_side, y_of_bound);
m_mpq_lar_core_solver.m_r_upper_bounds[j] = up;
}
set_upper_bound_witness(j, constr_ind);
break;
case GT:
y_of_bound = 1;
case GE:
m_mpq_lar_core_solver.m_column_types[j] = column_type::low_bound;
lean_assert(m_mpq_lar_core_solver.m_r_upper_bounds.size() > j);
{
auto low = numeric_pair<mpq>(right_side, y_of_bound);
m_mpq_lar_core_solver.m_r_low_bounds[j] = low;
}
set_low_bound_witness(j, constr_ind);
break;
case EQ:
m_mpq_lar_core_solver.m_column_types[j] = column_type::fixed;
m_mpq_lar_core_solver.m_r_low_bounds[j] = m_mpq_lar_core_solver.m_r_upper_bounds[j] = numeric_pair<mpq>(right_side, zero_of_type<mpq>());
set_upper_bound_witness(j, constr_ind);
set_low_bound_witness(j, constr_ind);
break;
default:
lean_unreachable();
}
m_columns_with_changed_bound.insert(j);
}
void update_upper_bound_column_type_and_bound(var_index j, lconstraint_kind kind, const mpq & right_side, constraint_index ci) {
lean_assert(m_mpq_lar_core_solver.m_column_types()[j] == column_type::upper_bound);
mpq y_of_bound(0);
switch (kind) {
case LT:
y_of_bound = -1;
case LE:
{
auto up = numeric_pair<mpq>(right_side, y_of_bound);
if (up < m_mpq_lar_core_solver.m_r_upper_bounds()[j]) {
m_mpq_lar_core_solver.m_r_upper_bounds[j] = up;
set_upper_bound_witness(j, ci);
m_columns_with_changed_bound.insert(j);
}
}
break;
case GT:
y_of_bound = 1;
case GE:
m_mpq_lar_core_solver.m_column_types[j] = column_type::boxed;
{
auto low = numeric_pair<mpq>(right_side, y_of_bound);
m_mpq_lar_core_solver.m_r_low_bounds[j] = low;
set_low_bound_witness(j, ci);
m_columns_with_changed_bound.insert(j);
if (low > m_mpq_lar_core_solver.m_r_upper_bounds[j]) {
m_status = INFEASIBLE;
m_infeasible_column_index = j;
} else {
m_mpq_lar_core_solver.m_column_types[j] = m_mpq_lar_core_solver.m_r_low_bounds()[j] < m_mpq_lar_core_solver.m_r_upper_bounds()[j]? column_type::boxed : column_type::fixed;
}
}
break;
case EQ:
{
auto v = numeric_pair<mpq>(right_side, zero_of_type<mpq>());
if (v > m_mpq_lar_core_solver.m_r_upper_bounds[j]) {
m_status = INFEASIBLE;
set_low_bound_witness(j, ci);
m_infeasible_column_index = j;
} else {
m_mpq_lar_core_solver.m_r_low_bounds[j] = m_mpq_lar_core_solver.m_r_upper_bounds[j] = v;
m_columns_with_changed_bound.insert(j);
set_low_bound_witness(j, ci);
set_upper_bound_witness(j, ci);
m_mpq_lar_core_solver.m_column_types[j] = column_type::fixed;
}
break;
}
break;
default:
lean_unreachable();
}
}
void update_boxed_column_type_and_bound(var_index j, lconstraint_kind kind, const mpq & right_side, constraint_index ci) {
lean_assert(m_status == INFEASIBLE || (m_mpq_lar_core_solver.m_column_types()[j] == column_type::boxed && m_mpq_lar_core_solver.m_r_low_bounds()[j] < m_mpq_lar_core_solver.m_r_upper_bounds()[j]));
mpq y_of_bound(0);
switch (kind) {
case LT:
y_of_bound = -1;
case LE:
{
auto up = numeric_pair<mpq>(right_side, y_of_bound);
if (up < m_mpq_lar_core_solver.m_r_upper_bounds[j]) {
m_mpq_lar_core_solver.m_r_upper_bounds[j] = up;
set_upper_bound_witness(j, ci);
m_columns_with_changed_bound.insert(j);
}
if (up < m_mpq_lar_core_solver.m_r_low_bounds[j]) {
m_status = INFEASIBLE;
lean_assert(false);
m_infeasible_column_index = j;
} else {
if (m_mpq_lar_core_solver.m_r_low_bounds()[j] == m_mpq_lar_core_solver.m_r_upper_bounds()[j])
m_mpq_lar_core_solver.m_column_types[j] = column_type::fixed;
}
}
break;
case GT:
y_of_bound = 1;
case GE:
{
auto low = numeric_pair<mpq>(right_side, y_of_bound);
if (low > m_mpq_lar_core_solver.m_r_low_bounds[j]) {
m_mpq_lar_core_solver.m_r_low_bounds[j] = low;
m_columns_with_changed_bound.insert(j);
set_low_bound_witness(j, ci);
}
if (low > m_mpq_lar_core_solver.m_r_upper_bounds[j]) {
m_status = INFEASIBLE;
m_infeasible_column_index = j;
} else if ( low == m_mpq_lar_core_solver.m_r_upper_bounds[j]) {
m_mpq_lar_core_solver.m_column_types[j] = column_type::fixed;
}
}
break;
case EQ:
{
auto v = numeric_pair<mpq>(right_side, zero_of_type<mpq>());
if (v < m_mpq_lar_core_solver.m_r_low_bounds[j]) {
m_status = INFEASIBLE;
m_infeasible_column_index = j;
set_upper_bound_witness(j, ci);
} else if (v > m_mpq_lar_core_solver.m_r_upper_bounds[j]) {
m_status = INFEASIBLE;
m_infeasible_column_index = j;
set_low_bound_witness(j, ci);
} else {
m_mpq_lar_core_solver.m_r_low_bounds[j] = m_mpq_lar_core_solver.m_r_upper_bounds[j] = v;
set_low_bound_witness(j, ci);
set_upper_bound_witness(j, ci);
m_mpq_lar_core_solver.m_column_types[j] = column_type::fixed;
m_columns_with_changed_bound.insert(j);
}
break;
}
default:
lean_unreachable();
}
}
void update_low_bound_column_type_and_bound(var_index j, lconstraint_kind kind, const mpq & right_side, constraint_index ci) {
lean_assert(m_mpq_lar_core_solver.m_column_types()[j] == column_type::low_bound);
mpq y_of_bound(0);
switch (kind) {
case LT:
y_of_bound = -1;
case LE:
{
auto up = numeric_pair<mpq>(right_side, y_of_bound);
m_mpq_lar_core_solver.m_r_upper_bounds[j] = up;
set_upper_bound_witness(j, ci);
m_columns_with_changed_bound.insert(j);
if (up < m_mpq_lar_core_solver.m_r_low_bounds[j]) {
m_status = INFEASIBLE;
m_infeasible_column_index = j;
} else {
m_mpq_lar_core_solver.m_column_types[j] = m_mpq_lar_core_solver.m_r_low_bounds()[j] < m_mpq_lar_core_solver.m_r_upper_bounds()[j]? column_type::boxed : column_type::fixed;
}
}
break;
case GT:
y_of_bound = 1;
case GE:
{
auto low = numeric_pair<mpq>(right_side, y_of_bound);
if (low > m_mpq_lar_core_solver.m_r_low_bounds[j]) {
m_mpq_lar_core_solver.m_r_low_bounds[j] = low;
m_columns_with_changed_bound.insert(j);
set_low_bound_witness(j, ci);
}
}
break;
case EQ:
{
auto v = numeric_pair<mpq>(right_side, zero_of_type<mpq>());
if (v < m_mpq_lar_core_solver.m_r_low_bounds[j]) {
m_status = INFEASIBLE;
m_infeasible_column_index = j;
set_upper_bound_witness(j, ci);
} else {
m_mpq_lar_core_solver.m_r_low_bounds[j] = m_mpq_lar_core_solver.m_r_upper_bounds[j] = v;
set_low_bound_witness(j, ci);
set_upper_bound_witness(j, ci);
m_mpq_lar_core_solver.m_column_types[j] = column_type::fixed;
}
m_columns_with_changed_bound.insert(j);
break;
}
default:
lean_unreachable();
}
}
void update_fixed_column_type_and_bound(var_index j, lconstraint_kind kind, const mpq & right_side, constraint_index ci) {
lean_assert(m_status == INFEASIBLE || (m_mpq_lar_core_solver.m_column_types()[j] == column_type::fixed && m_mpq_lar_core_solver.m_r_low_bounds()[j] == m_mpq_lar_core_solver.m_r_upper_bounds()[j]));
lean_assert(m_status == INFEASIBLE || (m_mpq_lar_core_solver.m_r_low_bounds()[j].y.is_zero() && m_mpq_lar_core_solver.m_r_upper_bounds()[j].y.is_zero()));
auto v = numeric_pair<mpq>(right_side, mpq(0));
mpq y_of_bound(0);
switch (kind) {
case LT:
if (v <= m_mpq_lar_core_solver.m_r_low_bounds[j]) {
m_status = INFEASIBLE;
m_infeasible_column_index = j;
set_upper_bound_witness(j, ci);
}
break;
case LE:
{
if (v < m_mpq_lar_core_solver.m_r_low_bounds[j]) {
m_status = INFEASIBLE;
m_infeasible_column_index = j;
set_upper_bound_witness(j, ci);
}
}
break;
case GT:
{
if (v >= m_mpq_lar_core_solver.m_r_upper_bounds[j]) {
m_status = INFEASIBLE;
m_infeasible_column_index =j;
set_low_bound_witness(j, ci);
}
}
break;
case GE:
{
if (v > m_mpq_lar_core_solver.m_r_upper_bounds[j]) {
m_status = INFEASIBLE;
m_infeasible_column_index = j;
set_low_bound_witness(j, ci);
}
}
break;
case EQ:
{
if (v < m_mpq_lar_core_solver.m_r_low_bounds[j]) {
m_status = INFEASIBLE;
m_infeasible_column_index = j;
set_upper_bound_witness(j, ci);
} else if (v > m_mpq_lar_core_solver.m_r_upper_bounds[j]) {
m_status = INFEASIBLE;
m_infeasible_column_index = j;
set_low_bound_witness(j, ci);
}
break;
}
default:
lean_unreachable();
}
}

2
src/util/lp/iterator_on_column.h
View file

@ -11,7 +11,7 @@ template <typename T, typename X>
struct iterator_on_column:linear_combination_iterator<T> {
const vector<column_cell>& m_column; // the offset in term coeffs
const static_matrix<T, X> & m_A;
int m_i = -1; // the initial offset in the column
int m_i; // the initial offset in the column
unsigned size() const { return m_column.size(); }
iterator_on_column(const vector<column_cell>& column, const static_matrix<T,X> & A) // the offset in term coeffs
:

7
src/util/lp/iterator_on_indexed_vector.h
View file

@ -8,8 +8,11 @@ namespace lean {
template <typename T>
struct iterator_on_indexed_vector:linear_combination_iterator<T> {
const indexed_vector<T> & m_v;
unsigned m_offset = 0;
iterator_on_indexed_vector(const indexed_vector<T> & v) : m_v(v){}
unsigned m_offset;
iterator_on_indexed_vector(const indexed_vector<T> & v) :
m_v(v),
m_offset(0)
{}
unsigned size() const { return m_v.m_index.size(); }
bool next(T & a, unsigned & i) {
if (m_offset >= m_v.m_index.size())

6
src/util/lp/iterator_on_pivot_row.h
View file

@ -7,12 +7,14 @@
namespace lean {
template <typename T>
struct iterator_on_pivot_row:linear_combination_iterator<T> {
bool m_basis_returned = false;
bool m_basis_returned;
const indexed_vector<T> & m_v;
unsigned m_basis_j;
iterator_on_indexed_vector<T> m_it;
unsigned size() const { return m_it.size(); }
iterator_on_pivot_row(const indexed_vector<T> & v, unsigned basis_j) : m_v(v), m_basis_j(basis_j), m_it(v) {}
iterator_on_pivot_row(const indexed_vector<T> & v, unsigned basis_j) :
m_basis_returned(false),
m_v(v), m_basis_j(basis_j), m_it(v) {}
bool next(T & a, unsigned & i) {
if (m_basis_returned == false) {
m_basis_returned = true;

4
src/util/lp/iterator_on_row.h
View file

@ -8,8 +8,8 @@ namespace lean {
template <typename T>
struct iterator_on_row:linear_combination_iterator<T> {
const vector<row_cell<T>> & m_row;
unsigned m_i= 0; // offset
iterator_on_row(const vector<row_cell<T>> & row) : m_row(row)
unsigned m_i; // offset
iterator_on_row(const vector<row_cell<T>> & row) : m_row(row), m_i(0)
{}
unsigned size() const { return m_row.size(); }
bool next(T & a, unsigned & i) {

9
src/util/lp/iterator_on_term_with_basis_var.h
View file

@ -8,14 +8,15 @@
#include "util/lp/lar_term.h"
namespace lean {
struct iterator_on_term_with_basis_var:linear_combination_iterator<mpq> {
std::unordered_map<unsigned, mpq>::const_iterator m_i; // the offset in term coeffs
bool m_term_j_returned = false;
const lar_term & m_term;
unsigned m_term_j;
std::unordered_map<unsigned, mpq>::const_iterator m_i; // the offset in term coeffs
bool m_term_j_returned;
unsigned m_term_j;
unsigned size() const {return static_cast<unsigned>(m_term.m_coeffs.size() + 1);}
iterator_on_term_with_basis_var(const lar_term & t, unsigned term_j) :
m_i(t.m_coeffs.begin()),
m_term(t),
m_i(t.m_coeffs.begin()),
m_term_j_returned(false),
m_term_j(term_j) {}
bool next(mpq & a, unsigned & i) {

19
src/util/lp/lar_core_solver.h
View file

@ -25,7 +25,7 @@ class lar_core_solver {
// to grow and is set to -1 otherwise
int m_sign_of_entering_delta;
vector<std::pair<mpq, unsigned>> m_infeasible_linear_combination;
int m_infeasible_sum_sign = 0; // todo: get rid of this field
int m_infeasible_sum_sign; // todo: get rid of this field
vector<numeric_pair<mpq>> m_right_sides_dummy;
vector<mpq> m_costs_dummy;
vector<double> m_d_right_sides_dummy;
@ -216,8 +216,6 @@ public:
void pop(unsigned k) {
m_stacked_simplex_strategy.pop(k);
bool use_tableau = m_stacked_simplex_strategy() != simplex_strategy_enum::no_tableau;
// rationals
if (!settings().use_tableau())
m_r_A.pop(k);
@ -232,7 +230,7 @@ public:
m_r_x.resize(m_r_A.column_count());
m_r_solver.m_costs.resize(m_r_A.column_count());
m_r_solver.m_d.resize(m_r_A.column_count());
if(!use_tableau)
if(!settings().use_tableau())
pop_markowitz_counts(k);
m_d_A.pop(k);
if (m_d_solver.m_factorization != nullptr) {
@ -242,13 +240,14 @@ public:
m_d_x.resize(m_d_A.column_count());
pop_basis(k);
m_stacked_simplex_strategy.pop(k);
settings().simplex_strategy() = m_stacked_simplex_strategy;
lean_assert(m_r_solver.basis_heading_is_correct());
lean_assert(!need_to_presolve_with_double_solver() || m_d_solver.basis_heading_is_correct());
}
bool need_to_presolve_with_double_solver() const {
return settings().presolve_with_double_solver_for_lar && !settings().use_tableau();
return settings().simplex_strategy() == simplex_strategy_enum::lu;
}
template <typename L>
@ -368,7 +367,7 @@ public:
s.m_x[j] = s.m_low_bounds[j];
break;
case column_type::boxed:
if (my_random() % 2) {
if (settings().random_next() % 2) {
s.m_x[j] = s.m_low_bounds[j];
} else {
s.m_x[j] = s.m_upper_bounds[j];
@ -600,7 +599,7 @@ public:
}
if (no_r_lu()) { // it is the case where m_d_solver gives a degenerated basis, we need to roll back
std::cout << "no_r_lu" << std::endl;
// std::cout << "no_r_lu" << std::endl;
catch_up_in_lu_in_reverse(changes_of_basis, m_r_solver);
m_r_solver.find_feasible_solution();
m_d_basis = m_r_basis;
@ -774,8 +773,8 @@ public:
}
mpq find_delta_for_strict_bounds() const{
mpq delta = numeric_traits<mpq>::one();
mpq find_delta_for_strict_bounds(const mpq & initial_delta) const{
mpq delta = initial_delta;
for (unsigned j = 0; j < m_r_A.column_count(); j++ ) {
if (low_bound_is_set(j))
update_delta(delta, m_r_low_bounds[j], m_r_x[j]);

3
src/util/lp/lar_core_solver.hpp
View file

@ -14,7 +14,8 @@ namespace lean {
lar_core_solver::lar_core_solver(
lp_settings & settings,
const column_namer & column_names
):
):
m_infeasible_sum_sign(0),
m_r_solver(m_r_A,
m_right_sides_dummy,
m_r_x,

735
src/util/lp/lar_solver.h
View file

@ -29,86 +29,38 @@
#include "util/lp/iterator_on_term_with_basis_var.h"
#include "util/lp/iterator_on_row.h"
#include "util/lp/quick_xplain.h"
#include "util/lp/conversion_helper.h"
namespace lean {
template <typename V>
struct conversion_helper {
static V get_low_bound(const column_info<mpq> & ci) {
return V(ci.get_low_bound(), ci.low_bound_is_strict()? 1 : 0);
}
static V get_upper_bound(const column_info<mpq> & ci) {
return V(ci.get_upper_bound(), ci.upper_bound_is_strict()? -1 : 0);
}
};
template<>
struct conversion_helper <double> {
static double get_upper_bound(const column_info<mpq> & ci) {
if (!ci.upper_bound_is_strict())
return ci.get_upper_bound().get_double();
double eps = 0.00001;
if (!ci.low_bound_is_set())
return ci.get_upper_bound().get_double() - eps;
eps = std::min((ci.get_upper_bound() - ci.get_low_bound()).get_double() / 1000, eps);
return ci.get_upper_bound().get_double() - eps;
}
static double get_low_bound(const column_info<mpq> & ci) {
if (!ci.low_bound_is_strict())
return ci.get_low_bound().get_double();
double eps = 0.00001;
if (!ci.upper_bound_is_set())
return ci.get_low_bound().get_double() + eps;
eps = std::min((ci.get_upper_bound() - ci.get_low_bound()).get_double() / 1000, eps);
return ci.get_low_bound().get_double() + eps;
}
};
struct constraint_index_and_column_struct {
int m_ci = -1;
int m_j = -1;
constraint_index_and_column_struct() {}
constraint_index_and_column_struct(unsigned ci, unsigned j):
m_ci(static_cast<int>(ci)),
m_j(static_cast<int>(j))
{}
bool operator==(const constraint_index_and_column_struct & a) const { return a.m_ci == m_ci && a.m_j == m_j; }
bool operator!=(const constraint_index_and_column_struct & a) const { return ! (*this == a);}
};
class lar_solver : public column_namer {
//////////////////// fields //////////////////////////
lp_settings m_settings;
stacked_value<lp_status> m_status = OPTIMAL;
std::unordered_map<unsigned, var_index> m_ext_vars_to_columns;
stacked_value<lp_status> m_status;
stacked_value<simplex_strategy_enum> m_simplex_strategy;
std::unordered_map<unsigned, var_index> m_ext_vars_to_columns;
vector<unsigned> m_columns_to_ext_vars_or_term_indices;
stacked_vector<ul_pair> m_vars_to_ul_pairs;
vector<lar_base_constraint*> m_constraints;
stacked_value<unsigned> m_constraint_count;
indexed_vector<mpq> m_incoming_buffer;
// the set of column indices j such that bounds have changed for j
int_set m_columns_with_changed_bound;
int_set m_rows_with_changed_bounds;
int_set m_basic_columns_with_changed_cost;
stacked_value<int> m_infeasible_column_index = -1; // such can be found at the initialization step
int_set m_basic_columns_with_changed_cost;
stacked_value<int> m_infeasible_column_index; // such can be found at the initialization step
stacked_value<unsigned> m_term_count;
public: // debug remove later
vector<lar_term*> m_terms;
private:
vector<lar_term*> m_orig_terms;
const var_index m_terms_start_index = 1000000;
const var_index m_terms_start_index;
indexed_vector<mpq> m_column_buffer;
std::function<column_type (unsigned)> m_column_type_function = [this] (unsigned j) {return m_mpq_lar_core_solver.m_column_types()[j];};
std::function<column_type (unsigned)> m_column_type_function;
public:
lar_core_solver m_mpq_lar_core_solver;
unsigned constraint_count() const {
return m_constraints.size();
}
const lar_base_constraint& get_constraint(unsigned ci) const {
return *(m_constraints[ci]);
}
unsigned constraint_count() const {
return m_constraints.size();
}
const lar_base_constraint& get_constraint(unsigned ci) const {
return *(m_constraints[ci]);
}
////////////////// methods ////////////////////////////////
static_matrix<mpq, numeric_pair<mpq>> & A_r() { return m_mpq_lar_core_solver.m_r_A;}
@ -128,10 +80,12 @@ public:
}
lar_solver() : m_mpq_lar_core_solver(
m_settings,
*this
) {}
lar_solver() : m_status(OPTIMAL),
m_infeasible_column_index(-1),
m_terms_start_index(1000000),
m_column_type_function ([this] (unsigned j) {return m_mpq_lar_core_solver.m_column_types()[j];}),
m_mpq_lar_core_solver(m_settings, *this)
{}
void set_propagate_bounds_on_pivoted_rows_mode(bool v) {
m_mpq_lar_core_solver.m_r_solver.m_pivoted_rows = v? (& m_rows_with_changed_bounds) : nullptr;
@ -146,25 +100,8 @@ public:
delete t;
}
var_index add_var(unsigned ext_j) {
var_index i;
lean_assert (ext_j < m_terms_start_index);
if (ext_j >= m_terms_start_index)
throw 0; // todo : what is the right was to exit?
if (try_get_val(m_ext_vars_to_columns, ext_j, i)) {
return i;
}
lean_assert(m_vars_to_ul_pairs.size() == A_r().column_count());
i = A_r().column_count();
m_vars_to_ul_pairs.push_back (ul_pair(static_cast<unsigned>(-1)));
register_new_ext_var_index(ext_j);
add_non_basic_var_to_core_fields();
lean_assert(sizes_are_correct());
return i;
}
#include "util/lp/init_lar_solver.h"
numeric_pair<mpq> const& get_value(var_index vi) const { return m_mpq_lar_core_solver.m_r_x[vi]; }
bool is_term(var_index j) const {
@ -177,98 +114,16 @@ public:
}
bool need_to_presolve_with_doubles() const { return m_mpq_lar_core_solver.need_to_presolve_with_double_solver(); }
void add_row_from_term_no_constraint(const lar_term * term) {
// j will be a new variable
unsigned j = A_r().column_count();
ul_pair ul(j);
m_vars_to_ul_pairs.push_back(ul);
add_basic_var_to_core_fields();
if (use_tableau()) {
auto it = iterator_on_term_with_basis_var(*term, j);
A_r().fill_last_row_with_pivoting(it,
m_mpq_lar_core_solver.m_r_solver.m_basis_heading);
m_mpq_lar_core_solver.m_r_solver.m_b.resize(A_r().column_count(), zero_of_type<mpq>());
} else {
fill_last_row_of_A_r(A_r(), term);
}
m_mpq_lar_core_solver.m_r_x[j] = get_basic_var_value_from_row_directly(A_r().row_count() - 1);
if (need_to_presolve_with_doubles())
fill_last_row_of_A_d(A_d(), term);
}
bool use_lu() const { return m_settings.simplex_strategy() == simplex_strategy_enum::lu; }
void add_constraint_from_term_and_create_new_column_row(unsigned term_j, const lar_term* term,
lconstraint_kind kind, const mpq & right_side) {
lean_assert(A_r().column_count() == m_mpq_lar_core_solver.m_r_solver.m_costs.size());
// j will be a new variable
unsigned j = A_r().column_count();
ul_pair ul(j);
m_vars_to_ul_pairs.push_back(ul);
add_basic_var_to_core_fields();
if (!m_settings.use_tableau()) {
fill_last_row_of_A_r(A_r(), term);
}
else {
auto it = iterator_on_term_with_basis_var(*term, j);
A_r().fill_last_row_with_pivoting(it,
m_mpq_lar_core_solver.m_r_solver.m_basis_heading);
m_mpq_lar_core_solver.m_r_solver.m_b.resize(A_r().column_count(), zero_of_type<mpq>());
}
m_mpq_lar_core_solver.m_r_x[A_r().column_count() - 1] = get_basic_var_value_from_row_directly(A_r().row_count() - 1);
fill_last_row_of_A_d(A_d(), term);
register_new_ext_var_index(term_j);
// m_constraints.size() is the index of the constrained that is about to be added
update_column_type_and_bound(j, kind, right_side - term->m_v, m_constraints.size());
m_constraints.push_back(new lar_term_constraint(term, kind, right_side));
lean_assert(A_r().column_count() == m_mpq_lar_core_solver.m_r_solver.m_costs.size());
}
void add_var_bound_on_constraint_for_term(var_index j, lconstraint_kind kind, const mpq & right_side, constraint_index ci) {
lean_assert(is_term(j));
unsigned adjusted_term_index = adjust_term_index(j);
unsigned term_j;
if (try_get_val(m_ext_vars_to_columns, j, term_j)) {
mpq rs = right_side - m_orig_terms[adjusted_term_index]->m_v;
m_constraints.push_back(new lar_term_constraint(m_orig_terms[adjusted_term_index], kind, right_side));
update_column_type_and_bound(term_j, kind, rs, ci);
}
else {
add_constraint_from_term_and_create_new_column_row(j, m_orig_terms[adjusted_term_index], kind, right_side);
}
}
void add_row_for_term(const lar_term * term) {
lean_assert(sizes_are_correct());
add_row_from_term_no_constraint(term);
lean_assert(sizes_are_correct());
}
bool sizes_are_correct() const {
lean_assert(!m_mpq_lar_core_solver.need_to_presolve_with_double_solver() || A_r().column_count() == A_d().column_count());
lean_assert(strategy_is_undecided() || !m_mpq_lar_core_solver.need_to_presolve_with_double_solver() || A_r().column_count() == A_d().column_count());
lean_assert(A_r().column_count() == m_mpq_lar_core_solver.m_r_solver.m_column_types.size());
lean_assert(A_r().column_count() == m_mpq_lar_core_solver.m_r_solver.m_costs.size());
lean_assert(A_r().column_count() == m_mpq_lar_core_solver.m_r_x.size());
return true;
}
constraint_index add_var_bound(var_index j, lconstraint_kind kind, const mpq & right_side) {
lean_assert(sizes_are_correct());
constraint_index ci = m_constraints.size();
if (!is_term(j)) { // j is a var
auto vc = new lar_var_constraint(j, kind, right_side);
m_constraints.push_back(vc);
update_column_type_and_bound(j, kind, right_side, ci);
} else {
add_var_bound_on_constraint_for_term(j, kind, right_side, ci);
}
lean_assert(sizes_are_correct());
return ci;
}
void print_implied_bound(const implied_bound& be, std::ostream & out) const {
out << "implied bound\n";
@ -490,10 +345,10 @@ public:
// new linear_combination_iterator_on_vector<mpq>(m_terms[adjust_term_index(term_index)]->coeffs_as_vector());
}
unsigned adjust_column_index_to_term_index(unsigned j) const {
unsigned ext_var_or_term = m_columns_to_ext_vars_or_term_indices[j];
return ext_var_or_term < m_terms_start_index ? j : ext_var_or_term;
}
unsigned adjust_column_index_to_term_index(unsigned j) const {
unsigned ext_var_or_term = m_columns_to_ext_vars_or_term_indices[j];
return ext_var_or_term < m_terms_start_index ? j : ext_var_or_term;
}
void propagate_bounds_on_a_term(const lar_term& t, bound_propagator & bp, unsigned term_offset) {
lean_assert(false); // not implemented
@ -561,6 +416,9 @@ public:
void set_status(lp_status s) {m_status = s;}
lp_status find_feasible_solution() {
if (strategy_is_undecided())
decide_on_strategy_and_adjust_initial_state();
m_mpq_lar_core_solver.m_r_solver.m_look_for_feasible_solution_only = true;
return solve();
}
@ -599,7 +457,8 @@ public:
return ret;
}
void push() {
lean_assert(sizes_are_correct());
m_simplex_strategy = m_settings.simplex_strategy();
m_simplex_strategy.push();
m_status.push();
m_vars_to_ul_pairs.push();
m_infeasible_column_index.push();
@ -608,7 +467,6 @@ public:
m_term_count.push();
m_constraint_count = m_constraints.size();
m_constraint_count.push();
lean_assert(sizes_are_correct());
}
static void clean_large_elements_after_pop(unsigned n, int_set& set) {
@ -627,8 +485,7 @@ public:
void pop(unsigned k) {
lean_assert(sizes_are_correct());
int n_was = static_cast<int>(m_ext_vars_to_columns.size());
int n_was = static_cast<int>(m_ext_vars_to_columns.size());
m_status.pop(k);
m_infeasible_column_index.pop(k);
unsigned n = m_vars_to_ul_pairs.peek_size(k);
@ -645,7 +502,8 @@ public:
unsigned m = A_r().row_count();
clean_large_elements_after_pop(m, m_rows_with_changed_bounds);
clean_inf_set_of_r_solver_after_pop();
lean_assert(!use_tableau() || m_mpq_lar_core_solver.m_r_solver.reduced_costs_are_correct_tableau());
lean_assert(m_settings.simplex_strategy() == simplex_strategy_enum::undecided ||
(!use_tableau()) || m_mpq_lar_core_solver.m_r_solver.reduced_costs_are_correct_tableau());
lean_assert(ax_is_correct());
@ -662,7 +520,9 @@ public:
}
m_terms.resize(m_term_count);
m_orig_terms.resize(m_term_count);
lean_assert(sizes_are_correct());
m_simplex_strategy.pop(k);
m_settings.simplex_strategy() = m_simplex_strategy;
lean_assert(sizes_are_correct());
lean_assert((!m_settings.use_tableau()) || m_mpq_lar_core_solver.m_r_solver.reduced_costs_are_correct_tableau());
}
@ -676,6 +536,9 @@ public:
bool maximize_term_on_tableau(const vector<std::pair<mpq, var_index>> & term,
impq &term_max) {
if (settings().simplex_strategy() == simplex_strategy_enum::undecided)
decide_on_strategy_and_adjust_initial_state();
m_mpq_lar_core_solver.solve();
if (m_mpq_lar_core_solver.m_r_solver.get_status() == UNBOUNDED)
return false;
@ -747,13 +610,13 @@ public:
bool maximize_term_on_corrected_r_solver(const vector<std::pair<mpq, var_index>> & term,
impq &term_max) {
settings().backup_costs = false;
switch (settings().m_simplex_strategy) {
switch (settings().simplex_strategy()) {
case simplex_strategy_enum::tableau_rows:
prepare_costs_for_r_solver(term);
settings().m_simplex_strategy = simplex_strategy_enum::tableau_costs;
settings().simplex_strategy() = simplex_strategy_enum::tableau_costs;
{
bool ret = maximize_term_on_tableau(term, term_max);
settings().m_simplex_strategy = simplex_strategy_enum::tableau_rows;
settings().simplex_strategy() = simplex_strategy_enum::tableau_rows;
set_costs_to_zero(term);
m_mpq_lar_core_solver.m_r_solver.set_status(OPTIMAL);
return ret;
@ -767,7 +630,7 @@ public:
return ret;
}
case simplex_strategy_enum::no_tableau:
case simplex_strategy_enum::lu:
lean_assert(false); // not implemented
return false;
default:
@ -786,24 +649,6 @@ public:
var_index add_term(const vector<std::pair<mpq, var_index>> & coeffs,
const mpq &m_v) {
lean_assert(A_r().column_count() == m_mpq_lar_core_solver.m_r_solver.m_costs.size());
m_terms.push_back(new lar_term(coeffs, m_v));
m_orig_terms.push_back(new lar_term(coeffs, m_v));
unsigned adjusted_term_index = m_terms.size() - 1;
if (use_tableau() && !coeffs.empty()) {
register_new_ext_var_index(m_terms_start_index + adjusted_term_index);
add_row_for_term(m_orig_terms.back());
if (m_settings.bound_propagation())
m_rows_with_changed_bounds.insert(A_r().row_count() - 1);
}
lean_assert(m_ext_vars_to_columns.size() == A_r().column_count());
lean_assert(A_r().column_count() == m_mpq_lar_core_solver.m_r_solver.m_costs.size());
return m_terms_start_index + adjusted_term_index;
}
const lar_term & get_term(unsigned j) const {
lean_assert(j >= m_terms_start_index);
return *m_terms[j - m_terms_start_index];
@ -818,78 +663,6 @@ public:
A_d().pop(k);
}
void add_new_var_to_core_fields_for_mpq(bool register_in_basis) {
unsigned j = A_r().column_count();
A_r().add_column();
lean_assert(m_mpq_lar_core_solver.m_r_x.size() == j);
// lean_assert(m_mpq_lar_core_solver.m_r_low_bounds.size() == j && m_mpq_lar_core_solver.m_r_upper_bounds.size() == j); // restore later
m_mpq_lar_core_solver.m_r_x.resize(j + 1);
m_mpq_lar_core_solver.m_r_low_bounds.increase_size_by_one();
m_mpq_lar_core_solver.m_r_upper_bounds.increase_size_by_one();
m_mpq_lar_core_solver.m_r_solver.m_inf_set.increase_size_by_one();
m_mpq_lar_core_solver.m_r_solver.m_costs.resize(j + 1);
m_mpq_lar_core_solver.m_r_solver.m_d.resize(j + 1);
lean_assert(m_mpq_lar_core_solver.m_r_heading.size() == j); // as A().column_count() on the entry to the method
if (register_in_basis) {
A_r().add_row();
m_mpq_lar_core_solver.m_r_heading.push_back(m_mpq_lar_core_solver.m_r_basis.size());
m_mpq_lar_core_solver.m_r_basis.push_back(j);
if (m_settings.bound_propagation())
m_rows_with_changed_bounds.insert(A_r().row_count() - 1);
} else {
m_mpq_lar_core_solver.m_r_heading.push_back(- static_cast<int>(m_mpq_lar_core_solver.m_r_nbasis.size()) - 1);
m_mpq_lar_core_solver.m_r_nbasis.push_back(j);
}
}
void add_new_var_to_core_fields_for_doubles(bool register_in_basis) {
unsigned j = A_d().column_count();
A_d().add_column();
lean_assert(m_mpq_lar_core_solver.m_d_x.size() == j);
// lean_assert(m_mpq_lar_core_solver.m_d_low_bounds.size() == j && m_mpq_lar_core_solver.m_d_upper_bounds.size() == j); // restore later
m_mpq_lar_core_solver.m_d_x.resize(j + 1 );
m_mpq_lar_core_solver.m_d_low_bounds.resize(j + 1);
m_mpq_lar_core_solver.m_d_upper_bounds.resize(j + 1);
lean_assert(m_mpq_lar_core_solver.m_d_heading.size() == j); // as A().column_count() on the entry to the method
if (register_in_basis) {
A_d().add_row();
m_mpq_lar_core_solver.m_d_heading.push_back(m_mpq_lar_core_solver.m_d_basis.size());
m_mpq_lar_core_solver.m_d_basis.push_back(j);
}else {
m_mpq_lar_core_solver.m_d_heading.push_back(- static_cast<int>(m_mpq_lar_core_solver.m_d_nbasis.size()) - 1);
m_mpq_lar_core_solver.m_d_nbasis.push_back(j);
}
}
void add_basic_var_to_core_fields() {
bool need_to_presolve_with_doubles = m_mpq_lar_core_solver.need_to_presolve_with_double_solver();
lean_assert(!need_to_presolve_with_doubles || A_r().column_count() == A_d().column_count());
m_mpq_lar_core_solver.m_column_types.push_back(column_type::free_column);
m_columns_with_changed_bound.increase_size_by_one();
m_rows_with_changed_bounds.increase_size_by_one();
add_new_var_to_core_fields_for_mpq(true);
if (need_to_presolve_with_doubles)
add_new_var_to_core_fields_for_doubles(true);
}
void add_non_basic_var_to_core_fields() {
lean_assert(!m_mpq_lar_core_solver.need_to_presolve_with_double_solver() || A_r().column_count() == A_d().column_count());
m_mpq_lar_core_solver.m_column_types.push_back(column_type::free_column);
m_columns_with_changed_bound.increase_size_by_one();
add_new_var_to_core_fields_for_mpq(false);
if (m_mpq_lar_core_solver.need_to_presolve_with_double_solver())
add_new_var_to_core_fields_for_doubles(false);
}
void register_new_ext_var_index(unsigned s) {
lean_assert(!contains(m_ext_vars_to_columns, s));
unsigned j = static_cast<unsigned>(m_ext_vars_to_columns.size());
m_ext_vars_to_columns[s] = j;
lean_assert(m_columns_to_ext_vars_or_term_indices.size() == j);
m_columns_to_ext_vars_or_term_indices.push_back(s);
}
void set_upper_bound_witness(var_index j, constraint_index ci) {
ul_pair ul = m_vars_to_ul_pairs[j];
@ -903,312 +676,6 @@ public:
m_vars_to_ul_pairs[j] = ul;
}
void update_free_column_type_and_bound(var_index j, lconstraint_kind kind, const mpq & right_side, constraint_index constr_ind) {
mpq y_of_bound(0);
switch (kind) {
case LT:
y_of_bound = -1;
case LE:
m_mpq_lar_core_solver.m_column_types[j] = column_type::upper_bound;
lean_assert(m_mpq_lar_core_solver.m_column_types()[j] == column_type::upper_bound);
lean_assert(m_mpq_lar_core_solver.m_r_upper_bounds.size() > j);
{
auto up = numeric_pair<mpq>(right_side, y_of_bound);
m_mpq_lar_core_solver.m_r_upper_bounds[j] = up;
}
set_upper_bound_witness(j, constr_ind);
break;
case GT:
y_of_bound = 1;
case GE:
m_mpq_lar_core_solver.m_column_types[j] = column_type::low_bound;
lean_assert(m_mpq_lar_core_solver.m_r_upper_bounds.size() > j);
{
auto low = numeric_pair<mpq>(right_side, y_of_bound);
m_mpq_lar_core_solver.m_r_low_bounds[j] = low;
}
set_low_bound_witness(j, constr_ind);
break;
case EQ:
m_mpq_lar_core_solver.m_column_types[j] = column_type::fixed;
m_mpq_lar_core_solver.m_r_low_bounds[j] = m_mpq_lar_core_solver.m_r_upper_bounds[j] = numeric_pair<mpq>(right_side, zero_of_type<mpq>());
set_upper_bound_witness(j, constr_ind);
set_low_bound_witness(j, constr_ind);
break;
default:
lean_unreachable();
}
m_columns_with_changed_bound.insert(j);
}
void update_upper_bound_column_type_and_bound(var_index j, lconstraint_kind kind, const mpq & right_side, constraint_index ci) {
lean_assert(m_mpq_lar_core_solver.m_column_types()[j] == column_type::upper_bound);
mpq y_of_bound(0);
switch (kind) {
case LT:
y_of_bound = -1;
case LE:
{
auto up = numeric_pair<mpq>(right_side, y_of_bound);
if (up < m_mpq_lar_core_solver.m_r_upper_bounds()[j]) {
m_mpq_lar_core_solver.m_r_upper_bounds[j] = up;
set_upper_bound_witness(j, ci);
m_columns_with_changed_bound.insert(j);
}
}
break;
case GT:
y_of_bound = 1;
case GE:
m_mpq_lar_core_solver.m_column_types[j] = column_type::boxed;
{
auto low = numeric_pair<mpq>(right_side, y_of_bound);
m_mpq_lar_core_solver.m_r_low_bounds[j] = low;
set_low_bound_witness(j, ci);
m_columns_with_changed_bound.insert(j);
if (low > m_mpq_lar_core_solver.m_r_upper_bounds[j]) {
m_status = INFEASIBLE;
m_infeasible_column_index = j;
} else {
m_mpq_lar_core_solver.m_column_types[j] = m_mpq_lar_core_solver.m_r_low_bounds()[j] < m_mpq_lar_core_solver.m_r_upper_bounds()[j]? column_type::boxed : column_type::fixed;
}
}
break;
case EQ:
{
auto v = numeric_pair<mpq>(right_side, zero_of_type<mpq>());
if (v > m_mpq_lar_core_solver.m_r_upper_bounds[j]) {
m_status = INFEASIBLE;
set_low_bound_witness(j, ci);
m_infeasible_column_index = j;
} else {
m_mpq_lar_core_solver.m_r_low_bounds[j] = m_mpq_lar_core_solver.m_r_upper_bounds[j] = v;
m_columns_with_changed_bound.insert(j);
set_low_bound_witness(j, ci);
set_upper_bound_witness(j, ci);
m_mpq_lar_core_solver.m_column_types[j] = column_type::fixed;
}
break;
}
break;
default:
lean_unreachable();
}
}
void update_boxed_column_type_and_bound(var_index j, lconstraint_kind kind, const mpq & right_side, constraint_index ci) {
lean_assert(m_status == INFEASIBLE || (m_mpq_lar_core_solver.m_column_types()[j] == column_type::boxed && m_mpq_lar_core_solver.m_r_low_bounds()[j] < m_mpq_lar_core_solver.m_r_upper_bounds()[j]));
mpq y_of_bound(0);
switch (kind) {
case LT:
y_of_bound = -1;
case LE:
{
auto up = numeric_pair<mpq>(right_side, y_of_bound);
if (up < m_mpq_lar_core_solver.m_r_upper_bounds[j]) {
m_mpq_lar_core_solver.m_r_upper_bounds[j] = up;
set_upper_bound_witness(j, ci);
m_columns_with_changed_bound.insert(j);
}
if (up < m_mpq_lar_core_solver.m_r_low_bounds[j]) {
m_status = INFEASIBLE;
lean_assert(false);
m_infeasible_column_index = j;
} else {
if (m_mpq_lar_core_solver.m_r_low_bounds()[j] == m_mpq_lar_core_solver.m_r_upper_bounds()[j])
m_mpq_lar_core_solver.m_column_types[j] = column_type::fixed;
}
}
break;
case GT:
y_of_bound = 1;
case GE:
{
auto low = numeric_pair<mpq>(right_side, y_of_bound);
if (low > m_mpq_lar_core_solver.m_r_low_bounds[j]) {
m_mpq_lar_core_solver.m_r_low_bounds[j] = low;
m_columns_with_changed_bound.insert(j);
set_low_bound_witness(j, ci);
}
if (low > m_mpq_lar_core_solver.m_r_upper_bounds[j]) {
m_status = INFEASIBLE;
m_infeasible_column_index = j;
} else if ( low == m_mpq_lar_core_solver.m_r_upper_bounds[j]) {
m_mpq_lar_core_solver.m_column_types[j] = column_type::fixed;
}
}
break;
case EQ:
{
auto v = numeric_pair<mpq>(right_side, zero_of_type<mpq>());
if (v < m_mpq_lar_core_solver.m_r_low_bounds[j]) {
m_status = INFEASIBLE;
m_infeasible_column_index = j;
set_upper_bound_witness(j, ci);
} else if (v > m_mpq_lar_core_solver.m_r_upper_bounds[j]) {
m_status = INFEASIBLE;
m_infeasible_column_index = j;
set_low_bound_witness(j, ci);
} else {
m_mpq_lar_core_solver.m_r_low_bounds[j] = m_mpq_lar_core_solver.m_r_upper_bounds[j] = v;
set_low_bound_witness(j, ci);
set_upper_bound_witness(j, ci);
m_mpq_lar_core_solver.m_column_types[j] = column_type::fixed;
m_columns_with_changed_bound.insert(j);
}
break;
}
default:
lean_unreachable();
}
}
void update_low_bound_column_type_and_bound(var_index j, lconstraint_kind kind, const mpq & right_side, constraint_index ci) {
lean_assert(m_mpq_lar_core_solver.m_column_types()[j] == column_type::low_bound);
mpq y_of_bound(0);
switch (kind) {
case LT:
y_of_bound = -1;
case LE:
{
auto up = numeric_pair<mpq>(right_side, y_of_bound);
m_mpq_lar_core_solver.m_r_upper_bounds[j] = up;
set_upper_bound_witness(j, ci);
m_columns_with_changed_bound.insert(j);
if (up < m_mpq_lar_core_solver.m_r_low_bounds[j]) {
m_status = INFEASIBLE;
m_infeasible_column_index = j;
} else {
m_mpq_lar_core_solver.m_column_types[j] = m_mpq_lar_core_solver.m_r_low_bounds()[j] < m_mpq_lar_core_solver.m_r_upper_bounds()[j]? column_type::boxed : column_type::fixed;
}
}
break;
case GT:
y_of_bound = 1;
case GE:
{
auto low = numeric_pair<mpq>(right_side, y_of_bound);
if (low > m_mpq_lar_core_solver.m_r_low_bounds[j]) {
m_mpq_lar_core_solver.m_r_low_bounds[j] = low;
m_columns_with_changed_bound.insert(j);
set_low_bound_witness(j, ci);
}
}
break;
case EQ:
{
auto v = numeric_pair<mpq>(right_side, zero_of_type<mpq>());
if (v < m_mpq_lar_core_solver.m_r_low_bounds[j]) {
m_status = INFEASIBLE;
m_infeasible_column_index = j;
set_upper_bound_witness(j, ci);
} else {
m_mpq_lar_core_solver.m_r_low_bounds[j] = m_mpq_lar_core_solver.m_r_upper_bounds[j] = v;
set_low_bound_witness(j, ci);
set_upper_bound_witness(j, ci);
m_mpq_lar_core_solver.m_column_types[j] = column_type::fixed;
}
m_columns_with_changed_bound.insert(j);
break;
}
default:
lean_unreachable();
}
}
void update_fixed_column_type_and_bound(var_index j, lconstraint_kind kind, const mpq & right_side, constraint_index ci) {
lean_assert(m_status == INFEASIBLE || (m_mpq_lar_core_solver.m_column_types()[j] == column_type::fixed && m_mpq_lar_core_solver.m_r_low_bounds()[j] == m_mpq_lar_core_solver.m_r_upper_bounds()[j]));
lean_assert(m_status == INFEASIBLE || (m_mpq_lar_core_solver.m_r_low_bounds()[j].y.is_zero() && m_mpq_lar_core_solver.m_r_upper_bounds()[j].y.is_zero()));
auto v = numeric_pair<mpq>(right_side, mpq(0));
mpq y_of_bound(0);
switch (kind) {
case LT:
if (v <= m_mpq_lar_core_solver.m_r_low_bounds[j]) {
m_status = INFEASIBLE;
m_infeasible_column_index = j;
set_upper_bound_witness(j, ci);
}
break;
case LE:
{
if (v < m_mpq_lar_core_solver.m_r_low_bounds[j]) {
m_status = INFEASIBLE;
m_infeasible_column_index = j;
set_upper_bound_witness(j, ci);
}
}
break;
case GT:
{
if (v >= m_mpq_lar_core_solver.m_r_upper_bounds[j]) {
m_status = INFEASIBLE;
m_infeasible_column_index =j;
set_low_bound_witness(j, ci);
}
}
break;
case GE:
{
if (v > m_mpq_lar_core_solver.m_r_upper_bounds[j]) {
m_status = INFEASIBLE;
m_infeasible_column_index = j;
set_low_bound_witness(j, ci);
}
}
break;
case EQ:
{
if (v < m_mpq_lar_core_solver.m_r_low_bounds[j]) {
m_status = INFEASIBLE;
m_infeasible_column_index = j;
set_upper_bound_witness(j, ci);
} else if (v > m_mpq_lar_core_solver.m_r_upper_bounds[j]) {
m_status = INFEASIBLE;
m_infeasible_column_index = j;
set_low_bound_witness(j, ci);
}
break;
}
default:
lean_unreachable();
}
}
void update_column_type_and_bound(var_index j, lconstraint_kind kind, const mpq & right_side, constraint_index constr_index) {
switch(m_mpq_lar_core_solver.m_column_types[j]) {
case column_type::free_column:
update_free_column_type_and_bound(j, kind, right_side, constr_index);
break;
case column_type::boxed:
update_boxed_column_type_and_bound(j, kind, right_side, constr_index);
break;
case column_type::low_bound:
update_low_bound_column_type_and_bound(j, kind, right_side, constr_index);
break;
case column_type::upper_bound:
update_upper_bound_column_type_and_bound(j, kind, right_side, constr_index);
break;
case column_type::fixed:
update_fixed_column_type_and_bound(j, kind, right_side, constr_index);
break;
default:
lean_assert(false); // cannot be here
}
}
void substitute_terms(const mpq & mult,
const vector<std::pair<mpq, var_index>>& left_side_with_terms,
@ -1247,8 +714,9 @@ public:
bool use_tableau() const { return m_settings.use_tableau(); }
bool use_tableau_costs() const { return m_settings.simplex_strategy() == simplex_strategy_enum::tableau_costs; }
bool use_tableau_costs() const {
return m_settings.simplex_strategy() == simplex_strategy_enum::tableau_costs;
}
void detect_rows_of_column_with_bound_change(unsigned j) {
if (m_mpq_lar_core_solver.m_r_heading[j] >= 0) { // it is a basic column
@ -1485,23 +953,6 @@ public:
unsigned basis_j = A.column_count() - 1;
A.set(last_row, basis_j, mpq(1));
}
// this fills the last row of A_d and sets the basis column: -1 in the last column of the row
void fill_last_row_of_A_d(static_matrix<double, double> & A, const lar_term* ls) {
lean_assert(A.row_count() > 0);
lean_assert(A.column_count() > 0);
unsigned last_row = A.row_count() - 1;
lean_assert(A.m_rows[last_row].empty());
for (auto & t : ls->m_coeffs) {
lean_assert(!is_zero(t.second));
var_index j = t.first;
A.set(last_row, j, - t.second.get_double());
}
unsigned basis_j = A.column_count() - 1;
A.set(last_row, basis_j, - 1 );
}
template <typename U, typename V>
void create_matrix_A(static_matrix<U, V> & matr) {
@ -1518,8 +969,8 @@ public:
template <typename U, typename V>
void copy_from_mpq_matrix(static_matrix<U, V> & matr) {
lean_assert(matr.row_count() == A_r().row_count());
lean_assert(matr.column_count() == A_r().column_count());
matr.m_rows.resize(A_r().row_count());
matr.m_columns.resize(A_r().column_count());
for (unsigned i = 0; i < matr.row_count(); i++) {
for (auto & it : A_r().m_rows[i]) {
matr.set(i, it.m_j, convert_struct<U, mpq>::convert(it.get_val()));
@ -1691,10 +1142,8 @@ public:
bool explanation_is_correct(const vector<std::pair<mpq, unsigned>>& explanation) const {
#ifdef LEAN_DEBUG
#if 0
// disabled as 'kind' is not assigned
lconstraint_kind kind;
the_relations_are_of_same_type(explanation, kind);
lean_assert(the_relations_are_of_same_type(explanation, kind));
lean_assert(the_left_sides_sum_to_zero(explanation));
mpq rs = sum_of_right_sides_of_explanation(explanation);
switch (kind) {
@ -1712,7 +1161,6 @@ public:
lean_assert(false);
return false;
}
#endif
#endif
return true;
}
@ -1737,58 +1185,6 @@ public:
return ret;
}
// template <typename U, typename V>
// void prepare_core_solver_fields_with_signature(static_matrix<U, V> & A, vector<V> & x,
// vector<V> & low_bound,
// vector<V> & upper_bound, const lar_solution_signature & signature) {
// create_matrix_A_r(A);
// fill_bounds_for_core_solver(low_bound, upper_bound);
// if (m_status == INFEASIBLE) {
// lean_assert(false); // not implemented
// }
// resize_and_init_x_with_signature(x, low_bound, upper_bound, signature);
// lean_assert(A.column_count() == x.size());
// }
// void find_solution_signature_with_doubles(lar_solution_signature & signature) {
// static_matrix<double, double> A;
// vector<double> x, low_bounds, upper_bounds;
// lean_assert(false); // not implemented
// // prepare_core_solver_fields<double, double>(A, x, low_bounds, upper_bounds);
// vector<double> column_scale_vector;
// vector<double> right_side_vector(A.row_count(), 0);
// scaler<double, double > scaler(right_side_vector,
// A,
// m_settings.scaling_minimum,
// m_settings.scaling_maximum,
// column_scale_vector,
// m_settings);
// if (!scaler.scale()) {
// // the scale did not succeed, unscaling
// A.clear();
// create_matrix_A_r(A);
// for (auto & s : column_scale_vector)
// s = 1.0;
// }
// vector<double> costs(A.column_count());
// auto core_solver = lp_primal_core_solver<double, double>(A,
// right_side_vector,
// x,
// m_mpq_lar_core_solver.m_basis,
// m_mpq_lar_core_solver.m_nbasis,
// m_mpq_lar_core_solver.m_heading,
// costs,
// m_mpq_lar_core_solver.m_column_types(),
// low_bounds,
// upper_bounds,
// m_settings,
// *this);
// core_solver.find_feasible_solution();
// extract_signature_from_lp_core_solver(core_solver, signature);
// }
bool has_lower_bound(var_index var, constraint_index& ci, mpq& value, bool& is_strict) {
if (var >= m_vars_to_ul_pairs.size()) {
@ -1865,12 +1261,29 @@ public:
void get_model(std::unordered_map<var_index, mpq> & variable_values) const {
mpq delta = mpq(1, 2); // start from 0.5 to have less clashes
lean_assert(m_status == OPTIMAL);
mpq delta = m_mpq_lar_core_solver.find_delta_for_strict_bounds();
for (unsigned i = 0; i < m_mpq_lar_core_solver.m_r_x.size(); i++ ) {
const numeric_pair<mpq> & rp = m_mpq_lar_core_solver.m_r_x[i];
variable_values[i] = rp.x + delta * rp.y;
}
unsigned i;
do {
// different pairs have to produce different singleton values
std::unordered_set<impq> set_of_different_pairs;
std::unordered_set<mpq> set_of_different_singles;
delta = m_mpq_lar_core_solver.find_delta_for_strict_bounds(delta);
for (i = 0; i < m_mpq_lar_core_solver.m_r_x.size(); i++ ) {
const numeric_pair<mpq> & rp = m_mpq_lar_core_solver.m_r_x[i];
set_of_different_pairs.insert(rp);
mpq x = rp.x + delta * rp.y;
set_of_different_singles.insert(x);
if (set_of_different_pairs.size()
!= set_of_different_singles.size()) {
delta /= mpq(2);
break;
}
variable_values[i] = x;
}
} while (i != m_mpq_lar_core_solver.m_r_x.size());
}

7
src/util/lp/linear_combination_iterator.h
View file

@ -16,7 +16,7 @@ struct linear_combination_iterator {
template <typename T>
struct linear_combination_iterator_on_vector : linear_combination_iterator<T> {
vector<std::pair<T, unsigned>> & m_vector;
int m_offset = 0;
int m_offset;
bool next(T & a, unsigned & i) {
if(m_offset >= m_vector.size())
return false;
@ -40,7 +40,10 @@ struct linear_combination_iterator_on_vector : linear_combination_iterator<T> {
linear_combination_iterator<T> * clone() {
return new linear_combination_iterator_on_vector(m_vector);
}
linear_combination_iterator_on_vector(vector<std::pair<T, unsigned>> & vec): m_vector(vec) {}
linear_combination_iterator_on_vector(vector<std::pair<T, unsigned>> & vec):
m_vector(vec),
m_offset(0)
{}
unsigned size() const { return m_vector.size(); }
};

71
src/util/lp/lp_core_solver_base.h
View file

@ -17,7 +17,8 @@ namespace lean {
template <typename T, typename X> // X represents the type of the x variable and the bounds
class lp_core_solver_base {
unsigned m_total_iterations = 0;
unsigned m_total_iterations;
unsigned m_iters_with_no_cost_growing;
unsigned inc_total_iterations() { ++m_settings.st().m_total_iterations; return m_total_iterations++; }
private:
lp_status m_status;
@ -25,40 +26,39 @@ public:
bool current_x_is_feasible() const { return m_inf_set.size() == 0; }
bool current_x_is_infeasible() const { return m_inf_set.size() != 0; }
int_set m_inf_set;
bool m_using_infeas_costs = false;
bool m_using_infeas_costs;
vector<unsigned> m_columns_nz; // m_columns_nz[i] keeps an approximate value of non zeroes the i-th column
vector<unsigned> m_rows_nz; // m_rows_nz[i] keeps an approximate value of non zeroes in the i-th row
indexed_vector<T> m_pivot_row_of_B_1; // the pivot row of the reverse of B
indexed_vector<T> m_pivot_row; // this is the real pivot row of the simplex tableu
vector<unsigned> m_columns_nz; // m_columns_nz[i] keeps an approximate value of non zeroes the i-th column
vector<unsigned> m_rows_nz; // m_rows_nz[i] keeps an approximate value of non zeroes in the i-th row
indexed_vector<T> m_pivot_row_of_B_1; // the pivot row of the reverse of B
indexed_vector<T> m_pivot_row; // this is the real pivot row of the simplex tableu
static_matrix<T, X> & m_A; // the matrix A
vector<X> & m_b; // the right side
vector<unsigned> & m_basis;
vector<unsigned>& m_nbasis;
vector<int>& m_basis_heading;
vector<X> & m_x; // a feasible solution, the fist time set in the constructor
vector<T> & m_costs;
lp_settings & m_settings;
vector<T> m_y; // the buffer for yB = cb
vector<X> & m_b; // the right side
vector<unsigned> & m_basis;
vector<unsigned>& m_nbasis;
vector<int>& m_basis_heading;
vector<X> & m_x; // a feasible solution, the fist time set in the constructor
vector<T> & m_costs;
lp_settings & m_settings;
vector<T> m_y; // the buffer for yB = cb
// a device that is able to solve Bx=c, xB=d, and change the basis
lu<T, X> * m_factorization = nullptr;
const column_namer & m_column_names;
indexed_vector<T> m_w; // the vector featuring in 24.3 of the Chvatal book
vector<T> m_d; // the vector of reduced costs
indexed_vector<T> m_ed; // the solution of B*m_ed = a
unsigned m_iters_with_no_cost_growing = 0;
lu<T, X> * m_factorization;
const column_namer & m_column_names;
indexed_vector<T> m_w; // the vector featuring in 24.3 of the Chvatal book
vector<T> m_d; // the vector of reduced costs
indexed_vector<T> m_ed; // the solution of B*m_ed = a
const vector<column_type> & m_column_types;
const vector<X> & m_low_bounds;
const vector<X> & m_upper_bounds;
vector<T> m_column_norms; // the approximate squares of column norms that help choosing a profitable column
vector<X> m_copy_of_xB;
unsigned m_basis_sort_counter = 0;
vector<T> m_steepest_edge_coefficients;
vector<unsigned> m_trace_of_basis_change_vector; // the even positions are entering, the odd positions are leaving
bool m_tracing_basis_changes = false;
int_set* m_pivoted_rows = nullptr;
bool m_look_for_feasible_solution_only = false;
const vector<X> & m_low_bounds;
const vector<X> & m_upper_bounds;
vector<T> m_column_norms; // the approximate squares of column norms that help choosing a profitable column
vector<X> m_copy_of_xB;
unsigned m_basis_sort_counter;
vector<T> m_steepest_edge_coefficients;
vector<unsigned> m_trace_of_basis_change_vector; // the even positions are entering, the odd positions are leaving
bool m_tracing_basis_changes;
int_set* m_pivoted_rows;
bool m_look_for_feasible_solution_only;
void start_tracing_basis_changes() {
m_trace_of_basis_change_vector.resize(0);
m_tracing_basis_changes = true;
@ -348,7 +348,7 @@ public:
if (x_is_at_bound(j))
break; // we should preserve x if possible
// snap randomly
if (my_random() % 2 == 1)
if (m_settings.random_next() % 2 == 1)
m_x[j] = m_low_bounds[j];
else
m_x[j] = m_upper_bounds[j];
@ -678,6 +678,13 @@ public:
lean_assert(is_zero(this->m_costs[j]));
}
return true;
}
}
unsigned & iters_with_no_cost_growing() {
return m_iters_with_no_cost_growing;
}
const unsigned & iters_with_no_cost_growing() const {
return m_iters_with_no_cost_growing;
}
};
}

26
src/util/lp/lp_core_solver_base.hpp
View file

@ -22,8 +22,11 @@ lp_core_solver_base(static_matrix<T, X> & A,
const vector<column_type> & column_types,
const vector<X> & low_bound_values,
const vector<X> & upper_bound_values):
m_total_iterations(0),
m_iters_with_no_cost_growing(0),
m_status(FEASIBLE),
m_inf_set(A.column_count()),
m_using_infeas_costs(false),
m_pivot_row_of_B_1(A.row_count()),
m_pivot_row(A.column_count()),
m_A(A),
@ -45,7 +48,11 @@ lp_core_solver_base(static_matrix<T, X> & A,
m_upper_bounds(upper_bound_values),
m_column_norms(m_n()),
m_copy_of_xB(m_m()),
m_steepest_edge_coefficients(A.column_count()) {
m_basis_sort_counter(0),
m_steepest_edge_coefficients(A.column_count()),
m_tracing_basis_changes(false),
m_pivoted_rows(nullptr),
m_look_for_feasible_solution_only(false) {
lean_assert(bounds_for_boxed_are_set_correctly());
init();
init_basis_heading_and_non_basic_columns_vector();
@ -57,10 +64,9 @@ allocate_basis_heading() { // the rest of initilization will be handled by the f
lean_assert(basis_heading_is_correct());
}
template <typename T, typename X> void lp_core_solver_base<T, X>::
init() {
my_random_init(m_settings.random_seed);
init() {
allocate_basis_heading();
if (!use_tableau())
if (m_settings.use_lu())
init_factorization(m_factorization, m_A, m_basis, m_settings);
}
@ -527,13 +533,19 @@ update_basis_and_x(int entering, int leaving, X const & tt) {
init_factorization(m_factorization, m_A, m_basis, m_settings);
if (!find_x_by_solving()) {
restore_x(entering, tt);
lean_assert(!A_mult_x_is_off());
if(A_mult_x_is_off()) {
m_status = FLOATING_POINT_ERROR;
m_iters_with_no_cost_growing++;
return false;
}
init_factorization(m_factorization, m_A, m_basis, m_settings);
m_iters_with_no_cost_growing++;
if (m_factorization->get_status() != LU_status::OK) {
std::stringstream s;
s << "failing refactor on off_result for entering = " << entering << ", leaving = " << leaving << " total_iterations = " << total_iterations();
throw_exception(s.str());
// s << "failing refactor on off_result for entering = " << entering << ", leaving = " << leaving << " total_iterations = " << total_iterations();
m_status = FLOATING_POINT_ERROR;
return false;
}
return false;
}

12
src/util/lp/lp_dual_core_solver.hpp
View file

@ -429,7 +429,7 @@ template <typename T, typename X> bool lp_dual_core_solver<T, X>::basis_change_a
if (snap_runaway_nonbasic_column(m_p)) {
if (!this->find_x_by_solving()) {
revert_to_previous_basis();
this->m_iters_with_no_cost_growing++;
this->iters_with_no_cost_growing()++;
return false;
}
}
@ -437,7 +437,7 @@ template <typename T, typename X> bool lp_dual_core_solver<T, X>::basis_change_a
if (!problem_is_dual_feasible()) {
// todo : shift the costs!!!!
revert_to_previous_basis();
this->m_iters_with_no_cost_growing++;
this->iters_with_no_cost_growing()++;
return false;
}
@ -537,7 +537,7 @@ template <typename T, typename X> unsigned lp_dual_core_solver<T, X>::get_number
if (this->m_m() > 300) {
s = (unsigned)((s / 100.0) * this->m_settings.percent_of_entering_to_check);
}
return my_random() % s + 1;
return this->m_settings.random_next() % s + 1;
}
template <typename T, typename X> bool lp_dual_core_solver<T, X>::delta_keeps_the_sign(int initial_delta_sign, const T & delta) {
@ -715,7 +715,7 @@ template <typename T, typename X> void lp_dual_core_solver<T, X>::update_xb_afte
template <typename T, typename X> void lp_dual_core_solver<T, X>::one_iteration() {
unsigned number_of_rows_to_try = get_number_of_rows_to_try_for_leaving();
unsigned offset_in_rows = my_random() % this->m_m();
unsigned offset_in_rows = this->m_settings.random_next() % this->m_m();
if (this->get_status() == TENTATIVE_DUAL_UNBOUNDED) {
number_of_rows_to_try = this->m_m();
} else {
@ -730,14 +730,14 @@ template <typename T, typename X> void lp_dual_core_solver<T, X>::solve() { // s
lean_assert(problem_is_dual_feasible());
lean_assert(this->basis_heading_is_correct());
this->set_total_iterations(0);
this->m_iters_with_no_cost_growing = 0;
this->iters_with_no_cost_growing() = 0;
do {
if (this->print_statistics_with_iterations_and_nonzeroes_and_cost_and_check_that_the_time_is_over("", *this->m_settings.get_message_ostream())){
return;
}
one_iteration();
} while (this->get_status() != FLOATING_POINT_ERROR && this->get_status() != DUAL_UNBOUNDED && this->get_status() != OPTIMAL &&
this->m_iters_with_no_cost_growing <= this->m_settings.max_number_of_iterations_with_no_improvements
this->iters_with_no_cost_growing() <= this->m_settings.max_number_of_iterations_with_no_improvements
&& this->total_iterations() <= this->m_settings.max_total_number_of_iterations);
}
}

3
src/util/lp/lp_dual_simplex.h
View file

@ -11,7 +11,7 @@ namespace lean {
template <typename T, typename X>
class lp_dual_simplex: public lp_solver<T, X> {
lp_dual_core_solver<T, X> * m_core_solver = nullptr;
lp_dual_core_solver<T, X> * m_core_solver;
vector<T> m_b_copy;
vector<T> m_low_bounds; // We don't have a convention here that all low bounds are zeros. At least it does not hold for the first stage solver
vector<column_type> m_column_types_of_core_solver;
@ -24,6 +24,7 @@ public:
}
}
lp_dual_simplex() : m_core_solver(nullptr) {}
void decide_on_status_after_stage1();

1
src/util/lp/lp_params.pyg
View file

@ -3,7 +3,6 @@ def_module_params('lp',
params=(
('rep_freq', UINT, 0, 'the report frequency, in how many iterations print the cost and other info '),
('min', BOOL, False, 'minimize cost'),
('presolve_with_dbl', BOOL, True, 'presolve with double'),
('print_stats', BOOL, False, 'print statistic'),
('simplex_strategy', UINT, 0, 'simplex strategy for the solver'),
('bprop_on_pivoted_rows', BOOL, True, 'propagate bounds on rows changed by the pivot operation')

40
src/util/lp/lp_primal_core_solver.h
View file

@ -32,21 +32,21 @@ class lp_primal_core_solver:public lp_core_solver_base<T, X> {
public:
// m_sign_of_entering is set to 1 if the entering variable needs
// to grow and is set to -1 otherwise
unsigned m_column_norm_update_counter;
T m_enter_price_eps;
int m_sign_of_entering_delta;
unsigned m_column_norm_update_counter;
T m_enter_price_eps;
int m_sign_of_entering_delta;
vector<breakpoint<X>> m_breakpoints;
binary_heap_priority_queue<X> m_breakpoint_indices_queue;
indexed_vector<T> m_beta; // see Swietanowski working vector beta for column norms
T m_epsilon_of_reduced_cost = T(1)/T(10000000);
vector<T> m_costs_backup;
T m_converted_harris_eps;
unsigned m_inf_row_index_for_tableau;
bool m_bland_mode_tableau;
int_set m_left_basis_tableau;
unsigned m_bland_mode_threshold = 1000;
unsigned m_left_basis_repeated;
vector<unsigned> m_leaving_candidates;
T m_epsilon_of_reduced_cost;
vector<T> m_costs_backup;
T m_converted_harris_eps;
unsigned m_inf_row_index_for_tableau;
bool m_bland_mode_tableau;
int_set m_left_basis_tableau;
unsigned m_bland_mode_threshold;
unsigned m_left_basis_repeated;
vector<unsigned> m_leaving_candidates;
// T m_converted_harris_eps = convert_struct<T, double>::convert(this->m_settings.harris_feasibility_tolerance);
std::list<unsigned> m_non_basis_list;
void sort_non_basis();
@ -76,10 +76,10 @@ public:
// choices.clear();
// choices.push_back(i);
// len = row_len;
// if (my_random() % 10) break;
// if (m_settings.random_next() % 10) break;
// } else if (row_len == len) {
// choices.push_back(i);
// if (my_random() % 10) break;
// if (m_settings.random_next() % 10) break;
// }
// }
@ -89,7 +89,7 @@ public:
// if (choices.size() == 1)
// return choices[0];
// unsigned k = my_random() % choices.size();
// unsigned k = this->m_settings.random_next() % choices.size();
// return choices[k];
// #endif
// }
@ -287,7 +287,7 @@ public:
choices.clear();
choices.push_back(&rc);
} else if (damage == num_of_non_free_basics &&
this->m_A.m_columns[j].size() <= len && (my_random() % 2)) {
this->m_A.m_columns[j].size() <= len && (this->m_settings.random_next() % 2)) {
choices.push_back(&rc);
len = this->m_A.m_columns[j].size();
}
@ -299,7 +299,7 @@ public:
return -1;
}
const row_cell<T>* rc = choices.size() == 1? choices[0] :
choices[my_random() % choices.size()];
choices[this->m_settings.random_next() % choices.size()];
a_ent = rc->m_value;
return rc->m_j;
@ -423,7 +423,7 @@ public:
void find_feasible_solution();
bool is_tiny() const {return this->m_m < 10 && this->m_n < 20;}
// bool is_tiny() const {return this->m_m < 10 && this->m_n < 20;}
void one_iteration();
void one_iteration_tableau();
@ -905,7 +905,9 @@ public:
column_type_array,
low_bound_values,
upper_bound_values),
m_beta(A.row_count()) {
m_beta(A.row_count()),
m_epsilon_of_reduced_cost(T(1)/T(10000000)),
m_bland_mode_threshold(1000) {
if (!(numeric_traits<T>::precise())) {
m_converted_harris_eps = convert_struct<T, double>::convert(this->m_settings.harris_feasibility_tolerance);

30
src/util/lp/lp_primal_core_solver.hpp
View file

@ -199,7 +199,7 @@ int lp_primal_core_solver<T, X>::choose_entering_column_presize(unsigned number_
entering_iter = non_basis_iter;
if (number_of_benefitial_columns_to_go_over)
number_of_benefitial_columns_to_go_over--;
} else if (t == j_nz && my_random() % 2 == 0) {
} else if (t == j_nz && this->m_settings.random_next() % 2 == 0) {
entering_iter = non_basis_iter;
}
}// while (number_of_benefitial_columns_to_go_over && initial_offset_in_non_basis != offset_in_nb);
@ -268,7 +268,7 @@ template <typename T, typename X> int lp_primal_core_solver<T, X>::advance_on_so
if (slope_at_entering * m_sign_of_entering_delta > - m_epsilon_of_reduced_cost) { // the slope started to increase infeasibility
break;
} else {
if ((numeric_traits<T>::precise() == false) || ( numeric_traits<T>::is_zero(slope_at_entering) && my_random() % 2 == 0)) {
if ((numeric_traits<T>::precise() == false) || ( numeric_traits<T>::is_zero(slope_at_entering) && this->m_settings.random_next() % 2 == 0)) {
// it is not cost benefitial to advance the delta more, so just break to increas the randomness
break;
}
@ -307,7 +307,7 @@ find_leaving_on_harris_theta(X const & harris_theta, X & t) {
// we also know that harris_theta is limited, so we will find a leaving
zero_harris_eps();
unsigned steps = this->m_ed.m_index.size();
unsigned k = my_random() % steps;
unsigned k = this->m_settings.random_next() % steps;
unsigned initial_k = k;
do {
unsigned i = this->m_ed.m_index[k];
@ -398,7 +398,7 @@ template <typename T, typename X> int lp_primal_core_solver<T, X>::find_leaving_
return find_leaving_and_t_with_breakpoints(entering, t);
bool unlimited = true;
unsigned steps = this->m_ed.m_index.size();
unsigned k = my_random() % steps;
unsigned k = this->m_settings.random_next() % steps;
unsigned initial_k = k;
unsigned row_min_nz = this->m_n() + 1;
m_leaving_candidates.clear();
@ -454,7 +454,7 @@ template <typename T, typename X> int lp_primal_core_solver<T, X>::find_leaving_
t = ratio;
return entering;
}
k = my_random() % m_leaving_candidates.size();
k = this->m_settings.random_next() % m_leaving_candidates.size();
return m_leaving_candidates[k];
}
@ -628,7 +628,7 @@ template <typename T, typename X> void lp_primal_core_solver<T, X>::backup_an
template <typename T, typename X> void lp_primal_core_solver<T, X>::init_run() {
this->m_basis_sort_counter = 0; // to initiate the sort of the basis
this->set_total_iterations(0);
this->m_iters_with_no_cost_growing = 0;
this->iters_with_no_cost_growing() = 0;
init_inf_set();
if (this->current_x_is_feasible() && this->m_look_for_feasible_solution_only)
return;
@ -664,7 +664,7 @@ void lp_primal_core_solver<T, X>::advance_on_entering_equal_leaving(int entering
this->init_lu();
if (!this->find_x_by_solving()) {
this->restore_x(entering, t * m_sign_of_entering_delta);
this->m_iters_with_no_cost_growing++;
this->iters_with_no_cost_growing()++;
LP_OUT(this->m_settings, "failing in advance_on_entering_equal_leaving for entering = " << entering << std::endl);
return;
}
@ -679,7 +679,7 @@ void lp_primal_core_solver<T, X>::advance_on_entering_equal_leaving(int entering
if (need_to_switch_costs() ||!this->current_x_is_feasible()) {
init_reduced_costs();
}
this->m_iters_with_no_cost_growing = 0;
this->iters_with_no_cost_growing() = 0;
}
template <typename T, typename X>void lp_primal_core_solver<T, X>::advance_on_entering_and_leaving(int entering, int leaving, X & t) {
@ -699,14 +699,14 @@ template <typename T, typename X>void lp_primal_core_solver<T, X>::advance_on_en
if (!pivot_compare_result){;}
else if (pivot_compare_result == 2) { // the sign is changed, cannot continue
this->set_status(UNSTABLE);
this->m_iters_with_no_cost_growing++;
this->iters_with_no_cost_growing()++;
return;
} else {
lean_assert(pivot_compare_result == 1);
this->init_lu();
if (this->m_factorization == nullptr || this->m_factorization->get_status() != LU_status::OK) {
this->set_status(UNSTABLE);
this->m_iters_with_no_cost_growing++;
this->iters_with_no_cost_growing()++;
return;
}
}
@ -728,7 +728,7 @@ template <typename T, typename X>void lp_primal_core_solver<T, X>::advance_on_en
}
if (!is_zero(t)) {
this->m_iters_with_no_cost_growing = 0;
this->iters_with_no_cost_growing() = 0;
init_infeasibility_after_update_x_if_inf(leaving);
}
@ -783,7 +783,7 @@ template <typename T, typename X> void lp_primal_core_solver<T, X>::advance_on_e
this->init_lu();
init_reduced_costs();
if (refresh_result == 2) {
this->m_iters_with_no_cost_growing++;
this->iters_with_no_cost_growing()++;
return;
}
}
@ -833,7 +833,7 @@ template <typename T, typename X> unsigned lp_primal_core_solver<T, X>::get_num
if (ret == 0) {
return 0;
}
return std::max(static_cast<unsigned>(my_random() % ret), 1u);
return std::max(static_cast<unsigned>(this->m_settings.random_next() % ret), 1u);
}
template <typename T, typename X> void lp_primal_core_solver<T, X>::print_column_norms(std::ostream & out) {
@ -934,7 +934,7 @@ template <typename T, typename X> unsigned lp_primal_core_solver<T, X>::solve()
&&
this->get_status() != INFEASIBLE
&&
this->m_iters_with_no_cost_growing <= this->m_settings.max_number_of_iterations_with_no_improvements
this->iters_with_no_cost_growing() <= this->m_settings.max_number_of_iterations_with_no_improvements
&&
this->total_iterations() <= this->m_settings.max_total_number_of_iterations
&&
@ -961,7 +961,7 @@ template <typename T, typename X> void lp_primal_core_solver<T, X>::init_column_
for (unsigned j = 0; j < this->m_n(); j++) {
this->m_column_norms[j] = T(static_cast<int>(this->m_A.m_columns[j].size() + 1))
+ T(static_cast<int>(my_random() % 10000)) / T(100000);
+ T(static_cast<int>(this->m_settings.random_next() % 10000)) / T(100000);
}
}

2
src/util/lp/lp_primal_core_solver_instances.cpp
View file

@ -11,6 +11,7 @@
#include "util/lp/lp_primal_core_solver.hpp"
#include "util/lp/lp_primal_core_solver_tableau.hpp"
namespace lean {
template void lp_primal_core_solver<double, double>::find_feasible_solution();
template void lean::lp_primal_core_solver<lean::mpq, lean::numeric_pair<lean::mpq> >::find_feasible_solution();
@ -22,4 +23,5 @@ template void lean::lp_primal_core_solver<double, double>::clear_breakpoints();
template bool lean::lp_primal_core_solver<lean::mpq, lean::mpq>::update_basis_and_x_tableau(int, int, lean::mpq const&);
template bool lean::lp_primal_core_solver<double, double>::update_basis_and_x_tableau(int, int, double const&);
template bool lean::lp_primal_core_solver<lean::mpq, lean::numeric_pair<lean::mpq> >::update_basis_and_x_tableau(int, int, lean::numeric_pair<lean::mpq> const&);
}

14
src/util/lp/lp_primal_core_solver_tableau.hpp
View file

@ -62,7 +62,7 @@ template <typename T, typename X> int lp_primal_core_solver<T, X>::choose_enteri
if (number_of_benefitial_columns_to_go_over)
number_of_benefitial_columns_to_go_over--;
}
else if (t == j_nz && my_random() % 2 == 0) {
else if (t == j_nz && this->m_settings.random_next() % 2 == 0) {
entering_iter = non_basis_iter;
}
}// while (number_of_benefitial_columns_to_go_over && initial_offset_in_non_basis != offset_in_nb);
@ -169,7 +169,7 @@ unsigned lp_primal_core_solver<T, X>::solve_with_tableau() {
&&
this->get_status() != INFEASIBLE
&&
this->m_iters_with_no_cost_growing <= this->m_settings.max_number_of_iterations_with_no_improvements
this->iters_with_no_cost_growing() <= this->m_settings.max_number_of_iterations_with_no_improvements
&&
this->total_iterations() <= this->m_settings.max_total_number_of_iterations
&&
@ -202,7 +202,7 @@ template <typename T, typename X>void lp_primal_core_solver<T, X>::advance_on_en
}
this->update_basis_and_x_tableau(entering, leaving, t);
lean_assert(this->A_mult_x_is_off() == false);
this->m_iters_with_no_cost_growing = 0;
this->iters_with_no_cost_growing() = 0;
} else {
this->pivot_column_tableau(entering, this->m_basis_heading[leaving]);
this->change_basis(entering, leaving);
@ -233,7 +233,7 @@ void lp_primal_core_solver<T, X>::advance_on_entering_equal_leaving_tableau(int
if (need_to_switch_costs()) {
init_reduced_costs_tableau();
}
this->m_iters_with_no_cost_growing = 0;
this->iters_with_no_cost_growing() = 0;
}
template <typename T, typename X> int lp_primal_core_solver<T, X>::find_leaving_and_t_tableau(unsigned entering, X & t) {
unsigned k = 0;
@ -293,7 +293,7 @@ template <typename T, typename X> int lp_primal_core_solver<T, X>::find_leaving_
}
if (m_leaving_candidates.size() == 1)
return m_leaving_candidates[0];
k = my_random() % m_leaving_candidates.size();
k = this->m_settings.random_next() % m_leaving_candidates.size();
return m_leaving_candidates[k];
}
template <typename T, typename X> void lp_primal_core_solver<T, X>::init_run_tableau() {
@ -302,7 +302,7 @@ template <typename T, typename X> void lp_primal_core_solver<T, X>::init_run_tab
lean_assert(basis_columns_are_set_correctly());
this->m_basis_sort_counter = 0; // to initiate the sort of the basis
this->set_total_iterations(0);
this->m_iters_with_no_cost_growing = 0;
this->iters_with_no_cost_growing() = 0;
lean_assert(this->inf_set_is_correct());
if (this->current_x_is_feasible() && this->m_look_for_feasible_solution_only)
return;
@ -315,7 +315,7 @@ template <typename T, typename X> void lp_primal_core_solver<T, X>::init_run_tab
this->m_column_norm_update_counter = 0;
init_column_norms();
}
if (this->m_settings.m_simplex_strategy == simplex_strategy_enum::tableau_rows)
if (this->m_settings.simplex_strategy() == simplex_strategy_enum::tableau_rows)
init_tableau_rows();
lean_assert(this->reduced_costs_are_correct_tableau());
lean_assert(!this->need_to_pivot_to_basis_tableau());

4
src/util/lp/lp_primal_simplex.h
View file

@ -15,7 +15,7 @@
namespace lean {
template <typename T, typename X>
class lp_primal_simplex: public lp_solver<T, X> {
lp_primal_core_solver<T, X> * m_core_solver = nullptr;
lp_primal_core_solver<T, X> * m_core_solver;
vector<X> m_low_bounds;
private:
unsigned original_rows() { return this->m_external_rows_to_core_solver_rows.size(); }
@ -28,7 +28,7 @@ private:
void set_scaled_costs();
public:
lp_primal_simplex() {}
lp_primal_simplex(): m_core_solver(nullptr) {}
column_info<T> * get_or_create_column_info(unsigned column);

141
src/util/lp/lp_settings.h
View file

@ -25,9 +25,10 @@ enum class column_type {
};
enum class simplex_strategy_enum {
undecided = 3,
tableau_rows = 0,
tableau_costs = 1,
no_tableau = 2
lu = 2
};
std::string column_type_to_string(column_type t);
@ -70,8 +71,6 @@ template <typename X> bool is_epsilon_small(const X & v, const double& eps);
int get_millisecond_count();
int get_millisecond_span(int start_time);
unsigned my_random();
void my_random_init(long unsigned seed);
class lp_resource_limit {
@ -105,49 +104,50 @@ private:
default_lp_resource_limit m_default_resource_limit;
lp_resource_limit* m_resource_limit;
// used for debug output
std::ostream* m_debug_out = &std::cout;
std::ostream* m_debug_out;
// used for messages, for example, the computation progress messages
std::ostream* m_message_out = &std::cout;
std::ostream* m_message_out;
stats m_stats;
random_gen m_rand;
public:
unsigned reps_in_scaler = 20;
unsigned reps_in_scaler;
// when the absolute value of an element is less than pivot_epsilon
// in pivoting, we treat it as a zero
double pivot_epsilon = 0.00000001;
double pivot_epsilon;
// see Chatal, page 115
double positive_price_epsilon = 1e-7;
double positive_price_epsilon;
// a quatation "if some choice of the entering vairable leads to an eta matrix
// whose diagonal element in the eta column is less than e2 (entering_diag_epsilon) in magnitude, the this choice is rejected ...
double entering_diag_epsilon = 1e-8;
int c_partial_pivoting = 10; // this is the constant c from page 410
unsigned depth_of_rook_search = 4;
bool using_partial_pivoting = true;
double entering_diag_epsilon;
int c_partial_pivoting; // this is the constant c from page 410
unsigned depth_of_rook_search;
bool using_partial_pivoting;
// dissertation of Achim Koberstein
// if Bx - b is different at any component more that refactor_epsilon then we refactor
double refactor_tolerance = 1e-4;
double pivot_tolerance = 1e-6;
double zero_tolerance = 1e-12;
double drop_tolerance = 1e-14;
double tolerance_for_artificials = 1e-4;
double can_be_taken_to_basis_tolerance = 0.00001;
double refactor_tolerance;
double pivot_tolerance;
double zero_tolerance;
double drop_tolerance;
double tolerance_for_artificials;
double can_be_taken_to_basis_tolerance;
unsigned percent_of_entering_to_check = 5; // we try to find a profitable column in a percentage of the columns
bool use_scaling = true;
double scaling_maximum = 1;
double scaling_minimum = 0.5;
double harris_feasibility_tolerance = 1e-7; // page 179 of Istvan Maros
double ignore_epsilon_of_harris = 10e-5;
unsigned max_number_of_iterations_with_no_improvements = 2000000;
unsigned max_total_number_of_iterations = 20000000;
double time_limit = std::numeric_limits<double>::max(); // the maximum time limit of the total run time in seconds
unsigned percent_of_entering_to_check; // we try to find a profitable column in a percentage of the columns
bool use_scaling;
double scaling_maximum;
double scaling_minimum;
double harris_feasibility_tolerance; // page 179 of Istvan Maros
double ignore_epsilon_of_harris;
unsigned max_number_of_iterations_with_no_improvements;
unsigned max_total_number_of_iterations;
double time_limit; // the maximum time limit of the total run time in seconds
// dual section
double dual_feasibility_tolerance = 1e-7; // // page 71 of the PhD thesis of Achim Koberstein
double primal_feasibility_tolerance = 1e-7; // page 71 of the PhD thesis of Achim Koberstein
double relative_primal_feasibility_tolerance = 1e-9; // page 71 of the PhD thesis of Achim Koberstein
double dual_feasibility_tolerance; // // page 71 of the PhD thesis of Achim Koberstein
double primal_feasibility_tolerance; // page 71 of the PhD thesis of Achim Koberstein
double relative_primal_feasibility_tolerance; // page 71 of the PhD thesis of Achim Koberstein
bool m_bound_propagation = true;
bool m_bound_propagation;
bool bound_progation() const {
return m_bound_propagation;
@ -157,7 +157,52 @@ public:
return m_bound_propagation;
}
lp_settings() : m_default_resource_limit(*this), m_resource_limit(&m_default_resource_limit) {}
lp_settings() : m_default_resource_limit(*this),
m_resource_limit(&m_default_resource_limit),
m_debug_out( &std::cout),
m_message_out(&std::cout),
reps_in_scaler(20),
pivot_epsilon(0.00000001),
positive_price_epsilon(1e-7),
entering_diag_epsilon ( 1e-8),
c_partial_pivoting ( 10), // this is the constant c from page 410
depth_of_rook_search ( 4),
using_partial_pivoting ( true),
// dissertation of Achim Koberstein
// if Bx - b is different at any component more that refactor_epsilon then we refactor
refactor_tolerance ( 1e-4),
pivot_tolerance ( 1e-6),
zero_tolerance ( 1e-12),
drop_tolerance ( 1e-14),
tolerance_for_artificials ( 1e-4),
can_be_taken_to_basis_tolerance ( 0.00001),
percent_of_entering_to_check ( 5),// we try to find a profitable column in a percentage of the columns
use_scaling ( true),
scaling_maximum ( 1),
scaling_minimum ( 0.5),
harris_feasibility_tolerance ( 1e-7), // page 179 of Istvan Maros
ignore_epsilon_of_harris ( 10e-5),
max_number_of_iterations_with_no_improvements ( 2000000),
max_total_number_of_iterations ( 20000000),
time_limit ( std::numeric_limits<double>::max()), // the maximum time limit of the total run time in seconds
// dual section
dual_feasibility_tolerance ( 1e-7), // // page 71 of the PhD thesis of Achim Koberstein
primal_feasibility_tolerance ( 1e-7), // page 71 of the PhD thesis of Achim Koberstein
relative_primal_feasibility_tolerance ( 1e-9), // page 71 of the PhD thesis of Achim Koberstein
m_bound_propagation ( true),
presolve_with_double_solver_for_lar(true),
m_simplex_strategy(simplex_strategy_enum::tableau_rows),
report_frequency(1000),
print_statistics(false),
column_norms_update_frequency(12000),
scale_with_ratio(true),
density_threshold(0.7),
use_breakpoints_in_feasibility_search(false),
max_row_length_for_bound_propagation(300),
backup_costs(true),
column_number_threshold_for_using_lu_in_lar_solver(4000)
{}
void set_resource_limit(lp_resource_limit& lim) { m_resource_limit = &lim; }
bool get_cancel_flag() const { return m_resource_limit->get_cancel_flag(); }
@ -226,8 +271,8 @@ public:
return is_eps_small_general<T>(t, tolerance_for_artificials);
}
// the method of lar solver to use
bool presolve_with_double_solver_for_lar = true;
simplex_strategy_enum m_simplex_strategy = simplex_strategy_enum::tableau_rows;
bool presolve_with_double_solver_for_lar;
simplex_strategy_enum m_simplex_strategy;
simplex_strategy_enum simplex_strategy() const {
return m_simplex_strategy;
}
@ -236,27 +281,33 @@ public:
return m_simplex_strategy;
}
bool use_lu() const {
return m_simplex_strategy == simplex_strategy_enum::lu;
}
bool use_tableau() const {
return m_simplex_strategy != simplex_strategy_enum::no_tableau;
return m_simplex_strategy == simplex_strategy_enum::tableau_rows ||
m_simplex_strategy == simplex_strategy_enum::tableau_costs;
}
bool use_tableau_rows() const {
return m_simplex_strategy == simplex_strategy_enum::tableau_rows;
}
int report_frequency = 1000;
bool print_statistics = false;
unsigned column_norms_update_frequency = 12000;
bool scale_with_ratio = true;
double density_threshold = 0.7; // need to tune it up, todo
int report_frequency;
bool print_statistics;
unsigned column_norms_update_frequency;
bool scale_with_ratio;
double density_threshold; // need to tune it up, todo
#ifdef LEAN_DEBUG
static unsigned ddd; // used for debugging
#endif
bool use_breakpoints_in_feasibility_search = false;
unsigned random_seed = 1;
static unsigned long random_next;
unsigned max_row_length_for_bound_propagation = 300;
bool backup_costs = true;
bool use_breakpoints_in_feasibility_search;
unsigned random_next() { return m_rand(); }
void random_seed(unsigned s) { m_rand.set_seed(s); }
unsigned max_row_length_for_bound_propagation;
bool backup_costs;
unsigned column_number_threshold_for_using_lu_in_lar_solver;
}; // end of lp_settings class

10
src/util/lp/lp_settings.hpp
View file

@ -67,15 +67,6 @@ int get_millisecond_span(int start_time) {
void my_random_init(long unsigned seed) {
lp_settings::random_next = seed;
}
unsigned my_random() {
lp_settings::random_next = lp_settings::random_next * 1103515245 + 12345;
return((unsigned)(lp_settings::random_next/65536) % 32768);
}
template <typename T>
bool vectors_are_equal(T * a, vector<T> &b, unsigned n) {
if (numeric_traits<T>::precise()) {
@ -126,7 +117,6 @@ bool vectors_are_equal(const vector<T> & a, const vector<T> &b) {
}
return true;
}
unsigned long lp_settings::random_next = 1;
#ifdef LEAN_DEBUG
unsigned lp_settings::ddd = 0;
#endif

29
src/util/lp/lp_solver.h
View file

@ -39,10 +39,10 @@ protected:
T get_column_cost_value(unsigned j, column_info<T> * ci) const;
public:
unsigned m_total_iterations;
static_matrix<T, X>* m_A = nullptr; // this is the matrix of constraints
static_matrix<T, X>* m_A; // this is the matrix of constraints
vector<T> m_b; // the right side vector
unsigned m_first_stage_iterations = 0;
unsigned m_second_stage_iterations = 0;
unsigned m_first_stage_iterations;
unsigned m_second_stage_iterations;
std::unordered_map<unsigned, lp_constraint<T, X>> m_constraints;
std::unordered_map<var_index, column_info<T>*> m_map_from_var_index_to_column_info;
std::unordered_map<unsigned, std::unordered_map<unsigned, T> > m_A_values;
@ -52,8 +52,8 @@ public:
std::unordered_map<unsigned, unsigned> m_core_solver_columns_to_external_columns;
vector<T> m_column_scale;
std::unordered_map<unsigned, std::string> m_name_map;
unsigned m_artificials = 0;
unsigned m_slacks = 0;
unsigned m_artificials;
unsigned m_slacks;
vector<column_type> m_column_types;
vector<T> m_costs;
vector<T> m_x;
@ -63,10 +63,17 @@ public:
vector<int> m_heading;
lp_status m_status = lp_status::UNKNOWN;
lp_status m_status;
lp_settings m_settings;
lp_solver() {}
lp_solver():
m_A(nullptr), // this is the matrix of constraints
m_first_stage_iterations (0),
m_second_stage_iterations (0),
m_artificials (0),
m_slacks (0),
m_status(lp_status::UNKNOWN)
{}
unsigned row_count() const { return this->m_A->row_count(); }
@ -232,14 +239,6 @@ protected:
out << "extended A[" << this->m_A->row_count() << "," << this->m_A->column_count() << "]" << std::endl;
}
struct row_tighten_stats {
unsigned n_of_new_bounds = 0;
unsigned n_of_fixed = 0;
bool is_obsolete = false;
};
public:
lp_settings & settings() { return m_settings;}
void print_model(std::ostream & s) const {

26
src/util/lp/lu.h
View file

@ -110,25 +110,25 @@ enum class LU_status { OK, Degenerated};
// Using Suhl-Suhl method described in the dissertation of Achim Koberstein, Chapter 5
template <typename T, typename X>
class lu {
LU_status m_status = LU_status::OK;
LU_status m_status;
public:
// the fields
unsigned m_dim;
unsigned m_dim;
static_matrix<T, X> const &m_A;
permutation_matrix<T, X> m_Q;
permutation_matrix<T, X> m_R;
permutation_matrix<T, X> m_r_wave;
sparse_matrix<T, X> m_U;
permutation_matrix<T, X> m_Q;
permutation_matrix<T, X> m_R;
permutation_matrix<T, X> m_r_wave;
sparse_matrix<T, X> m_U;
square_dense_submatrix<T, X>* m_dense_LU;
vector<tail_matrix<T, X> *> m_tail;
lp_settings & m_settings;
bool m_failure = false;
indexed_vector<T> m_row_eta_work_vector;
indexed_vector<T> m_w_for_extension;
indexed_vector<T> m_y_copy;
indexed_vector<unsigned> m_ii; //to optimize the work with the m_index fields
unsigned m_refactor_counter = 0;
lp_settings & m_settings;
bool m_failure;
indexed_vector<T> m_row_eta_work_vector;
indexed_vector<T> m_w_for_extension;
indexed_vector<T> m_y_copy;
indexed_vector<unsigned> m_ii; //to optimize the work with the m_index fields
unsigned m_refactor_counter;
// constructor
// if A is an m by n matrix then basis has length m and values in [0,n); the values are all different
// they represent the set of m columns

9
src/util/lp/lu.hpp
View file

@ -111,6 +111,7 @@ template <typename T, typename X>
lu<T, X>::lu(static_matrix<T, X> const & A,
vector<unsigned>& basis,
lp_settings & settings):
m_status(LU_status::OK),
m_dim(A.row_count()),
m_A(A),
m_Q(m_dim),
@ -118,7 +119,9 @@ lu<T, X>::lu(static_matrix<T, X> const & A,
m_r_wave(m_dim),
m_U(A, basis), // create the square matrix that eventually will be factorized
m_settings(settings),
m_row_eta_work_vector(A.row_count()){
m_failure(false),
m_row_eta_work_vector(A.row_count()),
m_refactor_counter(0) {
lean_assert(!(numeric_traits<T>::precise() && settings.use_tableau()));
#ifdef LEAN_DEBUG
debug_test_of_basis(A, basis);
@ -602,13 +605,13 @@ void lu<T, X>::process_column(int j) {
unsigned pi, pj;
bool success = m_U.get_pivot_for_column(pi, pj, m_settings.c_partial_pivoting, j);
if (!success) {
LP_OUT(m_settings, "get_pivot returned false: cannot find the pivot for column " << j << std::endl);
// LP_OUT(m_settings, "get_pivot returned false: cannot find the pivot for column " << j << std::endl);
m_failure = true;
return;
}
if (static_cast<int>(pi) == -1) {
LP_OUT(m_settings, "cannot find the pivot for column " << j << std::endl);
// LP_OUT(m_settings, "cannot find the pivot for column " << j << std::endl);
m_failure = true;
return;
}

56
src/util/lp/mps_reader.h
View file

@ -93,22 +93,28 @@ template <typename T, typename X>
class mps_reader {
enum row_type { Cost, Less_or_equal, Greater_or_equal, Equal };
struct bound {
bool m_low_is_set = true;
T m_low;
bool m_upper_is_set = false;
T m_upper;
bool m_value_is_fixed = false;
T m_fixed_value;
bool m_free = false;
T m_low;
T m_upper;
bool m_low_is_set;
bool m_upper_is_set;
bool m_value_is_fixed;
T m_fixed_value;
bool m_free;
// constructor
bound() : m_low(numeric_traits<T>::zero()) {} // it seems all mps files I have seen have the default low value 0 on a variable
bound() : m_low(numeric_traits<T>::zero()),
m_low_is_set(true),
m_upper_is_set(false),
m_value_is_fixed(false),
m_free(false) {} // it seems all mps files I have seen have the default low value 0 on a variable
};
struct column {
std::string m_name;
bound * m_bound = nullptr;
bound * m_bound;
unsigned m_index;
column(std::string name, unsigned index): m_name(name), m_index(index) {
column(std::string name, unsigned index): m_name(name),
m_bound(nullptr),
m_index(index) {
}
};
@ -116,15 +122,21 @@ class mps_reader {
row_type m_type;
std::string m_name;
std::unordered_map<std::string, T> m_row_columns;
T m_right_side = numeric_traits<T>::zero();
unsigned m_index;
T m_range = numeric_traits<T>::zero();
row(row_type type, std::string name, unsigned index) : m_type(type), m_name(name), m_index(index) {
T m_right_side;
T m_range;
row(row_type type, std::string name, unsigned index) :
m_type(type),
m_name(name),
m_index(index),
m_right_side(zero_of_type<T>()),
m_range(zero_of_type<T>())
{
}
};
bool m_is_OK;
std::string m_file_name;
bool m_is_OK = true;
std::unordered_map<std::string, row *> m_rows;
std::unordered_map<std::string, column *> m_columns;
std::unordered_map<std::string, unsigned> m_names_to_var_index;
@ -133,9 +145,9 @@ class mps_reader {
std::string m_cost_row_name;
std::ifstream m_file_stream;
// needed to adjust the index row
unsigned m_cost_line_count = 0;
unsigned m_line_number = 0;
std::ostream * m_message_stream = & std::cout;
unsigned m_cost_line_count;
unsigned m_line_number;
std::ostream * m_message_stream;
void set_m_ok_to_false() {
*m_message_stream << "setting m_is_OK to false" << std::endl;
@ -737,8 +749,12 @@ public:
}
mps_reader(std::string file_name):
m_file_name(file_name), m_file_stream(file_name) {
}
m_is_OK(true),
m_file_name(file_name),
m_file_stream(file_name),
m_cost_line_count(0),
m_line_number(0),
m_message_stream(& std::cout) {}
void read() {
if (!m_file_stream.is_open()){
set_m_ok_to_false();
@ -784,7 +800,7 @@ public:
auto it = m_names_to_var_index.find(s);
if (it != m_names_to_var_index.end())
return it->second;
unsigned ret = m_names_to_var_index.size();
unsigned ret = static_cast<unsigned>(m_names_to_var_index.size());
m_names_to_var_index[s] = ret;
return ret;
}

6
src/util/lp/numeric_pair.h
View file

@ -101,16 +101,14 @@ struct numeric_pair {
numeric_pair(T xp, T yp) : x(xp), y(yp) {}
template <typename X>
numeric_pair(const X & n) : x(n), y(0) {
}
template <typename X>
numeric_pair(const numeric_pair<X> & n) : x(n.x), y(n.y) {}
numeric_pair(const numeric_pair<T> & n) : x(n.x), y(n.y) {}
template <typename X, typename Y>
numeric_pair(X xp, Y yp) : numeric_pair(convert_struct<T, X>::convert(xp), convert_struct<T, Y>::convert(yp)) {}
numeric_pair(X xp, Y yp) : x(convert_struct<T, X>::convert(xp)), y(convert_struct<T, Y>::convert(yp)) {}
bool operator<(const numeric_pair& a) const {
return x < a.x || (x == a.x && y < a.y);

38
src/util/lp/permutation_matrix.h
View file

@ -132,42 +132,4 @@ class permutation_matrix : public tail_matrix<T, X> {
}; // end of the permutation class
#ifdef LEAN_DEBUG
template <typename T, typename X>
class permutation_generator {
unsigned m_n;
permutation_generator* m_lower;
bool m_done = false;
permutation_matrix<T, X> m_current;
unsigned m_last;
public:
permutation_generator(unsigned n);
permutation_generator(const permutation_generator & o);
bool move_next();
~permutation_generator() {
if (m_lower != nullptr) {
delete m_lower;
}
}
permutation_matrix<T, X> *current() {
return &m_current;
}
};
template <typename T, typename X>
inline unsigned number_of_inversions(permutation_matrix<T, X> & p);
template <typename T, typename X>
int sign(permutation_matrix<T, X> & p) {
return is_even(number_of_inversions(p))? 1: -1;
}
template <typename T, typename X>
T det_val_on_perm(permutation_matrix<T, X>* u, const matrix<T, X>& m);
template <typename T, typename X>
T determinant(const matrix<T, X>& m);
#endif
}

96
src/util/lp/permutation_matrix.hpp
View file

@ -320,100 +320,4 @@ template <typename T, typename X> bool permutation_matrix<T, X>::is_identity() c
}
#ifdef LEAN_DEBUG
template <typename T, typename X>
permutation_generator<T, X>::permutation_generator(unsigned n): m_n(n), m_current(n) {
lean_assert(n > 0);
if (n > 1) {
m_lower = new permutation_generator(n - 1);
} else {
m_lower = nullptr;
}
m_last = 0;
}
template <typename T, typename X>
permutation_generator<T, X>::permutation_generator(const permutation_generator & o): m_n(o.m_n), m_done(o.m_done), m_current(o.m_current), m_last(o.m_last) {
if (m_lower != nullptr) {
m_lower = new permutation_generator(o.m_lower);
} else {
m_lower = nullptr;
}
}
template <typename T, typename X> bool
permutation_generator<T, X>::move_next() {
if (m_done) {
return false;
}
if (m_lower == nullptr) {
if (m_last == 0) {
m_last++;
return true;
} else {
m_done = true;
return false;
}
} else {
if (m_last < m_n && m_last > 0) {
m_current[m_last - 1] = m_current[m_last];
m_current[m_last] = m_n - 1;
m_last++;
return true;
} else {
if (m_lower -> move_next()) {
auto lower_curr = m_lower -> current();
for ( unsigned i = 1; i < m_n; i++ ){
m_current[i] = (*lower_curr)[i - 1];
}
m_current[0] = m_n - 1;
m_last = 1;
return true;
} else {
m_done = true;
return false;
}
}
}
}
template <typename T, typename X>
inline unsigned number_of_inversions(permutation_matrix<T, X> & p) {
unsigned ret = 0;
unsigned n = p.size();
for (unsigned i = 0; i < n; i++) {
for (unsigned j = i + 1; j < n; j++) {
if (p[i] > p[j]) {
ret++;
}
}
}
return ret;
}
template <typename T, typename X>
T det_val_on_perm(permutation_matrix<T, X>* u, const matrix<T, X>& m) {
unsigned n = m.row_count();
T ret = numeric_traits<T>::one();
for (unsigned i = 0; i < n; i++) {
unsigned j = (*u)[i];
ret *= m(i, j);
}
return ret * sign(*u);
}
template <typename T, typename X>
T determinant(const matrix<T, X>& m) {
lean_assert(m.column_count() == m.row_count());
unsigned n = m.row_count();
permutation_generator<T, X> allp(n);
T ret = numeric_traits<T>::zero();
while (allp.move_next()){
ret += det_val_on_perm(allp.current(), m);
}
return ret;
}
#endif
}

5
src/util/lp/permutation_matrix_instances.cpp
View file

@ -46,11 +46,6 @@ template void lean::permutation_matrix<lean::mpq, lean::numeric_pair<lean::mpq>
template void lean::permutation_matrix<lean::mpq, lean::numeric_pair<lean::mpq> >::apply_reverse_from_left_to_T(vector<lean::mpq>&);
template void lean::permutation_matrix<lean::mpq, lean::numeric_pair<lean::mpq> >::apply_reverse_from_right_to_T(vector<lean::mpq >&);
template void lean::permutation_matrix<double, double>::multiply_by_permutation_from_right(lean::permutation_matrix<double, double>&);
#ifdef LEAN_DEBUG
template bool lean::permutation_generator<double, double>::move_next();
template lean::permutation_generator<double, double>::permutation_generator(unsigned int);
#endif
template lean::permutation_matrix<double, double>::permutation_matrix(unsigned int);
template void lean::permutation_matrix<double, double>::apply_reverse_from_left_to_X(vector<double> &);
template void lean::permutation_matrix< lean::mpq, lean::mpq>::apply_reverse_from_left_to_X(vector<lean::mpq> &);

9
src/util/lp/random_updater.h
View file

@ -16,12 +16,14 @@ namespace lean {
template <typename T> struct numeric_pair; // forward definition
class lar_core_solver; // forward definition
class random_updater {
unsigned range = 100000;
struct interval {
bool upper_bound_is_set = false;
bool upper_bound_is_set;
numeric_pair<mpq> upper_bound;
bool low_bound_is_set = false;
bool low_bound_is_set;
numeric_pair<mpq> low_bound;
interval() : upper_bound_is_set(false),
low_bound_is_set(false) {}
void set_low_bound(const numeric_pair<mpq> & v) {
if (low_bound_is_set) {
low_bound = std::max(v, low_bound);
@ -58,6 +60,7 @@ class random_updater {
};
std::set<var_index> m_var_set;
lar_core_solver & m_core_solver;
unsigned range;
linear_combination_iterator<mpq>* m_column_j; // the actual column
interval find_shift_interval(unsigned j);
interval get_interval_of_non_basic_var(unsigned j);

6
src/util/lp/random_updater.hpp
View file

@ -12,7 +12,9 @@ namespace lean {
random_updater::random_updater(
lar_core_solver & lar_core_solver,
const vector<unsigned> & column_indices) : m_core_solver(lar_core_solver) {
const vector<unsigned> & column_indices) :
m_core_solver(lar_core_solver),
range(100000) {
for (unsigned j : column_indices)
add_column_to_sets(j);
}
@ -134,7 +136,7 @@ void random_updater::shift_var(unsigned j, interval & r) {
}
numeric_pair<mpq> random_updater::get_random_from_interval(interval & r) {
unsigned rand = my_random();
unsigned rand = m_core_solver.settings().random_next();
if ((!r.low_bound_is_set) && (!r.upper_bound_is_set))
return numeric_pair<mpq>(rand % range, 0);
if (r.low_bound_is_set && (!r.upper_bound_is_set))

16
src/util/lp/sparse_matrix.h
View file

@ -30,10 +30,10 @@ class sparse_matrix
#endif
{
struct col_header {
unsigned m_shortened_markovitz = 0;
unsigned m_shortened_markovitz;
vector<indexed_value<T>> m_values; // the actual column values
col_header() {}
col_header(): m_shortened_markovitz(0) {}
void shorten_markovich_by_one() {
m_shortened_markovitz++;
@ -44,17 +44,17 @@ class sparse_matrix
}
};
unsigned m_n_of_active_elems = 0;
unsigned m_n_of_active_elems;
binary_heap_upair_queue<unsigned> m_pivot_queue;
public:
vector<vector<indexed_value<T>>> m_rows;
vector<col_header> m_columns;
permutation_matrix<T, X> m_row_permutation;
permutation_matrix<T, X> m_column_permutation;
vector<col_header> m_columns;
permutation_matrix<T, X> m_row_permutation;
permutation_matrix<T, X> m_column_permutation;
// m_work_pivot_vector[j] = offset of elementh of j-th column in the row we are pivoting to
// if the column is not present then m_work_pivot_vector[j] is -1
vector<int> m_work_pivot_vector;
vector<bool> m_processed;
vector<int> m_work_pivot_vector;
vector<bool> m_processed;
unsigned get_n_of_active_elems() const { return m_n_of_active_elems; }
#ifdef LEAN_DEBUG

1
src/util/lp/sparse_matrix.hpp
View file

@ -36,6 +36,7 @@ void sparse_matrix<T, X>::copy_B(static_matrix<T, X> const &A, vector<unsigned>
// constructor that copies columns of the basis from A
template <typename T, typename X>
sparse_matrix<T, X>::sparse_matrix(static_matrix<T, X> const &A, vector<unsigned> & basis) :
m_n_of_active_elems(0),
m_pivot_queue(A.row_count()),
m_row_permutation(A.row_count()),
m_column_permutation(A.row_count()),

2
src/util/lp/sparse_vector.h
View file

@ -20,7 +20,7 @@ public:
}
#ifdef LEAN_DEBUG
T operator[] (unsigned i) const {
for (auto t : m_data) {
for (auto &t : m_data) {
if (t.first == i) return t.second;
}
return numeric_traits<T>::zero();

2
src/util/lp/square_dense_submatrix.h
View file

@ -42,7 +42,7 @@ public:
unsigned m_index_start;
unsigned m_dim;
vector<T> m_v;
sparse_matrix<T, X> * m_parent = nullptr;
sparse_matrix<T, X> * m_parent;
permutation_matrix<T, X> m_row_permutation;
indexed_vector<T> m_work_vector;
public:

1
src/util/lp/static_matrix.hpp
View file

@ -29,7 +29,6 @@ template <typename T, typename X> void static_matrix<T, X>::scan_row_ii_to_offse
template <typename T, typename X> bool static_matrix<T, X>::pivot_row_to_row_given_cell(unsigned i, column_cell & c, unsigned pivot_col) {
// std::cout << "ddd = " << ++lp_settings::ddd<< std::endl;
unsigned ii = c.m_i;
lean_assert(i < row_count() && ii < column_count());
lean_assert(i != ii);

17
src/util/lp/ul_pair.h
View file

@ -32,14 +32,14 @@ inline bool compare(const std::pair<mpq, var_index> & a, const std::pair<mpq, va
}
class ul_pair {
constraint_index m_low_bound_witness = static_cast<constraint_index>(-1);
constraint_index m_upper_bound_witness = static_cast<constraint_index>(-1);
constraint_index m_low_bound_witness;
constraint_index m_upper_bound_witness;
public:
constraint_index& low_bound_witness() {return m_low_bound_witness;}
constraint_index low_bound_witness() const {return m_low_bound_witness;}
constraint_index& upper_bound_witness() { return m_upper_bound_witness;}
constraint_index upper_bound_witness() const {return m_upper_bound_witness;}
row_index m_i = static_cast<row_index>(-1);
row_index m_i;
bool operator!=(const ul_pair & p) const {
return !(*this == p);
}
@ -50,8 +50,15 @@ public:
m_i == p.m_i;
}
// empty constructor
ul_pair(){}
ul_pair(row_index ri) : m_i(ri) {}
ul_pair() :
m_low_bound_witness(static_cast<constraint_index>(-1)),
m_upper_bound_witness(static_cast<constraint_index>(-1)),
m_i(static_cast<row_index>(-1))
{}
ul_pair(row_index ri) :
m_low_bound_witness(static_cast<constraint_index>(-1)),
m_upper_bound_witness(static_cast<constraint_index>(-1)),
m_i(ri) {}
ul_pair(const ul_pair & o): m_low_bound_witness(o.m_low_bound_witness), m_upper_bound_witness(o.m_upper_bound_witness), m_i(o.m_i) {}
};