namespace for Limbo.Solvers More...

Namespaces
	lpmcf
	namespace for Limbo.Solvers.lpmcf

Classes
class	CapacityScaling
	Capacity scaling algorithm for min-cost flow. More...

struct	CompareTermByVariable
	Comapre term by variable. More...

class	CostScaling
	Cost scaling algorithm for min-cost flow. More...

class	CycleCanceling
	Cycle canceling algorithm for min-cost flow. More...

class	DualMinCostFlow
	LP solved with min-cost flow. A better implementation of limbo::solvers::lpmcf::LpDualMcf. More...

class	FeasibleSearcher
	Base heuristic to search for feasible solutions. More...

class	GurobiLinearApi
	Gurobi API with limbo::solvers::LinearModel. More...

class	GurobiParameters
	Base class for custom Gurobi parameters. More...

class	L2NormScaler
	Scaling scheme with default L2 norm scaling. More...

class	LagMultiplierUpdater
	A base helper function object to update lagrangian multipliers using subgradient descent. All other schemes can be derived from this class. More...

class	LinearConstraint
	Describe linear constraint. More...

class	LinearExpression
	Describe linear expressions in optimization problem. More...

class	LinearModel
	model to describe an optimization problem More...

class	LinearTerm
	Linear term. More...

class	LPSolveLinearApi
	LPSolve API with limbo::solvers::LinearModel. More...

class	LPSolveParameters
	Base class for custom LPSolve parameters. More...

struct	MatrixCSR
	Compressed sparse row (CSR) matrix. More...

class	MinCoefficientScaler
	Scaling scheme with minimum coefficient in an expression. More...

class	MinCostFlow
	LP solved with min-cost flow. More...

class	MinCostFlowSolver
	A base class of min-cost flow solver. More...

class	MultiKnapsackLagRelax
	Solve multiple knapsack problem with lagrangian relaxation. More...

class	NetworkSimplex
	Network simplex algorithm for min-cost flow. More...

class	ProblemScaler
	Base class for scaling scheme with default no scaling. More...

class	SearchByAdjustCoefficient
	Heuristic to search for feasible solutions by adjusting coefficients so that some items will not be assigned to some bins. More...

class	SearchByBinSmoothing
	Heuristic to search for feasible solutions by smoothing dense bins. More...

class	SearchByCombinedStrategy
	Heuristic to search for feasible solutions by combined strategies. More...

class	SubGradientDescent
	Update lagrangian multiplier with subgradient descent. More...

class	Variable
	Describe variables in optimization problem. More...

class	VariableProperty
	Describe properties of a variable. More...

Enumerations
enum	SolverProperty { MIN, MAX, BINARY, INTEGER, CONTINUOUS, OPTIMAL, INFEASIBLE, SUBOPTIMAL, UNBOUNDED }
	Some enums used in solver. More...

Functions
template<typename T , typename V >
void	axpy (unsigned int n, T a, V const x, T y)
	\( y = a \cdot x+y \) More...

template<typename T , typename V , typename MatrixType >
void	AxPlusy (T a, MatrixType const &A, V const x, T y)
	\( y = a A x + y \) More...

template<typename T , typename V , typename MatrixType >
void	ATxPlusy (T a, MatrixType const &A, V const x, T y)
	\( y = a A^T x + y \) More...

template<typename T >
T	dot (unsigned int n, T const x, T const y)
	compute dot product \( x^T y \) More...

template<typename T >
void	vcopy (unsigned int n, T const x, T y)
	copy vector More...

std::string	toString (SolverProperty sp)
	Convert limbo::solvers::SolverProperty to std::string. More...

template<typename T >
int	easy_sdp_ext (int n, int k, struct blockmatrix C, double a, struct constraintmatrix constraints, double constant_offset, struct blockmatrix pX, double py, struct blockmatrix pZ, double ppobj, double pdobj, struct paramstruc const &params, int const &printlevel)
	API to call Csdp solver. More...

Detailed Description

namespace for Limbo.Solvers

namespace for Solvers

namespace Limbo.Solvers

Enumeration Type Documentation

enum limbo::solvers::SolverProperty

Some enums used in solver.

Enumerator
MIN	minimize objective
MAX	maximize objective
BINARY	binary number
INTEGER	integer number
CONTINUOUS	floating point number
OPTIMAL	optimally solved
INFEASIBLE	the model is infeasible
SUBOPTIMAL	the model is suboptimal
UNBOUNDED	the model is unbounded

Definition at line 29 of file Solvers.h.

Function Documentation

template<typename T , typename V , typename MatrixType >

void limbo::solvers::ATxPlusy	(	T	a,
		MatrixType const &	A,
		V const *	x,
		T *	y
	)

inline

\( y = a A^T x + y \)

Template Parameters

T	data type of a, y
V	data type of x
MatrixType	sparse matrix type in CSR format

Parameters

a	constant
A	matrix
x	vector
y	output vector

Definition at line 226 of file Numerical.h.

template<typename T , typename V , typename MatrixType >

void limbo::solvers::AxPlusy	(	T	a,
		MatrixType const &	A,
		V const *	x,
		T *	y
	)

inline

\( y = a A x + y \)

Template Parameters

T	data type of a, y
V	data type of x
MatrixType	sparse matrix type in CSR format

Parameters

a	constant
A	matrix
x	vector
y	output vector

Definition at line 198 of file Numerical.h.

template<typename T , typename V >

void limbo::solvers::axpy	(	unsigned int	n,
		T	a,
		V const *	x,
		T *	y
	)

inline

\( y = a \cdot x+y \)

Template Parameters

T	data type of a, y
V	data type of x

Parameters

n	dimension
a	constant
x	vector
y	output vector

Definition at line 179 of file Numerical.h.

template<typename T >

T limbo::solvers::dot	(	unsigned int	n,
		T const *	x,
		T const *	y
	)

inline

compute dot product \( x^T y \)

Template Parameters

T data type

Parameters

n	dimension
x	vector
y	vector

Returns: dot product

Definition at line 252 of file Numerical.h.

template<typename T >

int limbo::solvers::easy_sdp_ext	(	int	n,
		int	k,
		struct blockmatrix	C,
		double *	a,
		struct constraintmatrix *	constraints,
		double	constant_offset,
		struct blockmatrix *	pX,
		double **	py,
		struct blockmatrix *	pZ,
		double *	ppobj,
		double *	pdobj,
		struct paramstruc const &	params,
		int const &	printlevel
	)

API to call Csdp solver.

This is a dummy template, so I do not need to write a .c file. See the documentation of Csdp for details on how to describe an SDP problem.

Template Parameters

T	dummy data type, any type works

Parameters

n,k,C,a,constraints,constant_offset	as input problem
pX,py,pZ	as initial solution and final solution
ppobj	as primal objective
pdobj	as dual objective
params	pass customized parameters to control the solver
printlevel	verbose level in printing

Returns: the return code from sdp

Definition at line 59 of file CsdpEasySdpApi.h.

std::string limbo::solvers::toString ( SolverProperty sp )

inline

Convert limbo::solvers::SolverProperty to std::string.

Parameters

sp	solver property

Definition at line 44 of file Solvers.h.

template<typename T >

void limbo::solvers::vcopy	(	unsigned int	n,
		T const *	x,
		T *	y
	)

inline

copy vector

Template Parameters

T data type

Parameters

n	dimension
x	source vector
y	target vector

Definition at line 269 of file Numerical.h.

Namespaces

Classes

Enumerations

Functions

Detailed Description

Enumeration Type Documentation

Function Documentation