limbo::solvers::SearchByCombinedStrategy< T, V > Class Template Reference

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

#include <MultiKnapsackLagRelax.h>

Inheritance diagram for limbo::solvers::SearchByCombinedStrategy< T, V >:

Public Types
typedef FeasibleSearcher< T, V >	base_type
	base type
typedef base_type::model_type	model_type
	model type
typedef base_type::solver_type	solver_type
	solver type
typedef solver_type::updater_type	updater_type
	updater type for lagrangian multipliers
typedef model_type::coefficient_value_type	coefficient_value_type
	coefficient value type
typedef model_type::variable_value_type	variable_value_type
	variable value type
typedef model_type::variable_type	variable_type
	variable type
typedef model_type::expression_type	expression_type
	expression type
typedef model_type::constraint_type	constraint_type
	constraint type
typedef model_type::term_type	term_type
	term type
Public Types inherited from limbo::solvers::FeasibleSearcher< T, V >
typedef LinearModel< T, V >	model_type
	model type
typedef MultiKnapsackLagRelax < T, V >	solver_type
	solver type
typedef solver_type::updater_type	updater_type
	updater type for lagrangian multipliers
typedef model_type::coefficient_value_type	coefficient_value_type
	coefficient value type
typedef model_type::variable_value_type	variable_value_type
	variable value type
typedef model_type::variable_type	variable_type
	variable type
typedef model_type::expression_type	expression_type
	expression type
typedef model_type::constraint_type	constraint_type
	constraint type
typedef model_type::term_type	term_type
	term type
typedef solver_type::matrix_type	matrix_type
	matrix type

Public Member Functions
	SearchByCombinedStrategy (solver_type *solver, coefficient_value_type convergeRatio=0.1)
	constructor More...
	~SearchByCombinedStrategy ()
	destructor
virtual SolverProperty	operator() (updater_type *updater)
	API to search for feasible solutions. More...
Public Member Functions inherited from limbo::solvers::FeasibleSearcher< T, V >
	FeasibleSearcher (solver_type *solver)
	constructor More...
virtual	~FeasibleSearcher ()
	destructor

Protected Attributes
SearchByAdjustCoefficient < coefficient_value_type, variable_value_type >	m_searcherCoeff
	search by adjusting coefficient
SearchByBinSmoothing < coefficient_value_type, variable_value_type >	m_searcherSmoothing
	search by smoothing dense bins
Protected Attributes inherited from limbo::solvers::FeasibleSearcher< T, V >
solver_type *	m_solver
	problem solver
model_type *const &	m_model
	model for the problem
coefficient_value_type *&	m_vObjCoef
	coefficients variables in objective
matrix_type const &	m_constrMatrix
	constraint matrix \(A\)
coefficient_value_type *const &	m_vConstrRhs
	constraint right hand side \(b\)
variable_type *const &	m_vGroupedVariable
	array of grouped variables according to item
unsigned int *const &	m_vVariableGroupBeginIndex
	begin index of grouped variable
unsigned int const &	m_numGroups
	number of groups
std::vector< unsigned int > const &	m_vConstraintPartition
	indices of constraints, the first partition is capacity constraints
coefficient_value_type *&	m_vLagMultiplier
	array of lagrangian multipliers
coefficient_value_type *&	m_vSlackness
	array of slackness values in each iteration, \( b-Ax \)
std::vector < coefficient_value_type > const &	m_vScalingFactor
	scaling factor for constraints and objective, last entry is for objective
coefficient_value_type &	m_objConstant
	constant value in objective from lagrangian relaxation
coefficient_value_type &	m_lagObj
	current objective of the lagrangian subproblem
unsigned int &	m_iter
	current iteration
unsigned int &	m_maxIters
	maximum number of iterations
bool &	m_useInitialSol
	whether use initial solutions or not
std::vector < variable_value_type > &	m_vBestVariableSol
	best feasible solution found so far
coefficient_value_type &	m_bestObj
	best objective found so far

Additional Inherited Members
Protected Member Functions inherited from limbo::solvers::FeasibleSearcher< T, V >
void	computeSlackness ()
	compute slackness in an iteration
SolverProperty	solveSubproblems (updater_type *updater, unsigned int beginIter, unsigned int endIter)
	kernel lagrangian iterations More...

Detailed Description

template<typename T, typename V>
class limbo::solvers::SearchByCombinedStrategy< T, V >

Heuristic to search for feasible solutions by combined strategies.

Template Parameters

T	coefficient value type
V	variable value type

Definition at line 1725 of file MultiKnapsackLagRelax.h.

Constructor & Destructor Documentation

template<typename T , typename V >

limbo::solvers::SearchByCombinedStrategy< T, V >::SearchByCombinedStrategy ( solver_type *  solver, coefficient_value_type  convergeRatio = 0.1  )

inline

constructor

Parameters

solver	problem solver
convergeRatio	ratio of convergence for limbo::solvers::SearchByAdjustCoefficient

Definition at line 1752 of file MultiKnapsackLagRelax.h.

Member Function Documentation

template<typename T , typename V >

virtual SolverProperty limbo::solvers::SearchByCombinedStrategy< T, V >::operator() ( updater_type *  updater )

inlinevirtual

API to search for feasible solutions.

Parameters

updater

updater for lagrangian multipliers

Returns

solving status

Reimplemented from limbo::solvers::FeasibleSearcher< T, V >.

Definition at line 1764 of file MultiKnapsackLagRelax.h.

---

The documentation for this class was generated from the following file:

• /Users/yibolin/Documents/Projects/Limbo/limbo/solvers/MultiKnapsackLagRelax.h