line::lp Namespace Reference

Classes
struct	LpSolution
class	LpModel
	Sparse LP in the natural form, with per-variable bounds. More...

Enumerations
enum class	LpStatus { Optimal , Infeasible , Unbounded , IterationLimit }
	Outcome of a solve. More...
enum class	LpSense { LE , EQ , GE }
	Row relation. More...

Functions
bool	highs_available ()
	True when a sparse backend is compiled in.
template<class T>
LpSolution< T >	lp_solve (const LpModel< T > &model, std::size_t dense_max_cols=512)
	Solve, choosing the backend by arithmetic and size.
const char *	lp_status_name (LpStatus s)
template<class T>
T	simplex_tolerance ()
template<class T>
LpSolution< T >	simplex_solve (const LpModel< T > &model, std::size_t max_iterations=0)
	Solve the model.

Enumeration Type Documentation

LpSense

enum class line::lp::LpSense

strong

Row relation.

Enumerator
LE
EQ
GE

Definition at line 103 of file simplex.h.

LpStatus

enum class line::lp::LpStatus

strong

Outcome of a solve.

Enumerator
Optimal	an optimal vertex was reached
Infeasible	phase 1 ended with residual artificial mass
Unbounded	an improving column has no blocking row
IterationLimit	the iteration cap was hit (cannot happen under Bland's rule with exact arithmetic; a guard for inexact T)

Definition at line 76 of file simplex.h.

Function Documentation

highs_available()

bool line::lp::highs_available ( )

inline

True when a sparse backend is compiled in.

Definition at line 51 of file lp_highs.h.

References highs_available().

Referenced by highs_available().

lp_solve()

template<class T>

LpSolution< T > line::lp::lp_solve	(	const LpModel< T > &	model,
		std::size_t	dense_max_cols = 512 )

Solve, choosing the backend by arithmetic and size.

Rational and every extended-precision T always take the exact dense path, at compile time, because that is where this tree's exactness guarantees live. A double model goes to HiGHS only once it is wider than dense_max_cols, so small models keep bit-for-bit the answers their goldens were taken with and the two backends stay comparable on exactly the instances the tests use.

Definition at line 176 of file lp_highs.h.

References lp_solve(), and simplex_solve().

Referenced by lp_solve(), line::mapqn::mapqn_qr_bounds_bas(), line::mapqn::mapqn_qr_bounds_rsrd(), line::npfqn::npfqn_bnd_bgt(), line::npfqn::npfqn_bnd_bpt(), line::mapqn::qrf_feasible_start_lp(), and line::mapqn::solve_qrf_nlp_lp().

lp_status_name()

const char * line::lp::lp_status_name ( LpStatus s )

inline

Definition at line 84 of file simplex.h.

References Infeasible, lp_status_name(), Optimal, and Unbounded.

Referenced by line::fj::fj_tsm_capacity(), lp_status_name(), line::mam::m3pp22_interleave_fitc(), line::mapqn::mapqn_bnd_lr(), line::mapqn::mapqn_bnd_lr_mva(), line::mapqn::mapqn_bnd_lr_pf(), line::mapqn::mapqn_bnd_qr(), line::mapqn::mapqn_qr_bounds_bas(), line::mapqn::mapqn_qr_bounds_rsrd(), line::npfqn::npfqn_bnd_bgt(), and line::npfqn::npfqn_bnd_bpt().

simplex_solve()

template<class T>

LpSolution< T > line::lp::simplex_solve	(	const LpModel< T > &	model,
		std::size_t	max_iterations = 0 )

Solve the model.

See the header comment for the bound handling; the returned x is in the caller's variable space and satisfies the bounds exactly when T is exact.

Definition at line 286 of file simplex.h.

References line::lp::LpModel< T >::col_at(), line::lp::LpModel< T >::costs(), EQ, GE, Infeasible, IterationLimit, line::lp::LpSolution< T >::iterations, line::lp::LpModel< T >::lower(), line::lp::LpModel< T >::lower_is_free(), line::lp::LpModel< T >::maximize(), line::lp::LpModel< T >::num_rows(), line::lp::LpModel< T >::num_vars(), line::lp::LpSolution< T >::objective, Optimal, line::lp::LpModel< T >::rhs(), line::lp::LpModel< T >::row_begin(), line::lp::LpModel< T >::row_end(), line::lp::LpModel< T >::sense(), simplex_solve(), simplex_tolerance(), line::lp::LpSolution< T >::status, Unbounded, line::lp::LpModel< T >::upper(), line::lp::LpModel< T >::upper_is_free(), line::lp::LpModel< T >::val_at(), and line::lp::LpSolution< T >::x.

Referenced by line::fj::fj_tsm_capacity(), lp_solve(), line::mam::m3pp22_interleave_fitc(), line::mapqn::mapqn_bnd_lr(), line::mapqn::mapqn_bnd_lr_mva(), line::mapqn::mapqn_bnd_lr_pf(), line::mapqn::mapqn_bnd_qr(), line::mapqn::qrf_feasible_start(), and simplex_solve().

simplex_tolerance()

template<class T>

T line::lp::simplex_tolerance ( )

inline

Definition at line 258 of file simplex.h.

References simplex_tolerance().

Referenced by simplex_solve(), and simplex_tolerance().