Ipopt::AmplTNLP Class Reference

Ampl Interface. More...

#include <AmplTNLP.hpp>

Inheritance diagram for Ipopt::AmplTNLP:

[legend]

Collaboration diagram for Ipopt::AmplTNLP:

[legend]

List of all members.

Public Member Functions
	DECLARE_STD_EXCEPTION (NONPOSITIVE_SCALING_FACTOR)
	Exceptions.
void	set_active_objective (Index obj_no)
	A method for setting the index of the objective function to be considered.
Constructors/Destructors

	AmplTNLP (const SmartPtr< const Journalist > &jnlst, const SmartPtr< OptionsList > options, char *&argv, SmartPtr< AmplSuffixHandler > suffix_handler=NULL, bool allow_discrete=false, SmartPtr< AmplOptionsList > ampl_options_list=NULL, const char ampl_option_string=NULL, const char ampl_invokation_string=NULL, const char ampl_banner_string=NULL, std::string *nl_file_content=NULL)
	Constructor.
virtual	~AmplTNLP ()
	Default destructor.
methods to gather information about the NLP. These
methods are overloaded from TNLP. See TNLP for their more detailed documentation.
virtual bool	get_nlp_info (Index &n, Index &m, Index &nnz_jac_g, Index &nnz_h_lag, IndexStyleEnum &index_style)
	returns dimensions of the nlp.
virtual bool	get_bounds_info (Index n, Number x_l, Number x_u, Index m, Number g_l, Number g_u)
	returns bounds of the nlp.
virtual bool	get_constraints_linearity (Index m, LinearityType *const_types)
	Returns the constraint linearity.
virtual bool	get_starting_point (Index n, bool init_x, Number x, bool init_z, Number z_L, Number z_U, Index m, bool init_lambda, Number lambda)
	provides a starting point for the nlp variables.
virtual bool	eval_f (Index n, const Number *x, bool new_x, Number &obj_value)
	evaluates the objective value for the nlp.
virtual bool	eval_grad_f (Index n, const Number x, bool new_x, Number grad_f)
	evaluates the gradient of the objective for the nlp.
virtual bool	eval_g (Index n, const Number x, bool new_x, Index m, Number g)
	evaluates the constraint residuals for the nlp.
virtual bool	eval_jac_g (Index n, const Number x, bool new_x, Index m, Index nele_jac, Index iRow, Index jCol, Number values)
	specifies the jacobian structure (if values is NULL) and evaluates the jacobian values (if values is not NULL) for the nlp.
virtual bool	eval_h (Index n, const Number x, bool new_x, Number obj_factor, Index m, const Number lambda, bool new_lambda, Index nele_hess, Index iRow, Index jCol, Number *values)
	specifies the structure of the hessian of the lagrangian (if values is NULL) and evaluates the values (if values is not NULL).
virtual bool	get_scaling_parameters (Number &obj_scaling, bool &use_x_scaling, Index n, Number x_scaling, bool &use_g_scaling, Index m, Number g_scaling)
	retrieve the scaling parameters for the variables, objective function, and constraints.
Solution Methods

virtual void	finalize_solution (SolverReturn status, Index n, const Number x, const Number z_L, const Number z_U, Index m, const Number g, const Number lambda, Number obj_value, const IpoptData ip_data, IpoptCalculatedQuantities *ip_cq)
	This method is called when the algorithm is complete so the TNLP can store/write the solution.
Method for quasi-Newton approximation information.

virtual Index	get_number_of_nonlinear_variables ()
virtual bool	get_list_of_nonlinear_variables (Index num_nonlin_vars, Index *pos_nonlin_vars)
Ampl specific methods

ASL_pfgh *	AmplSolverObject ()
	Return the ampl solver object (ASL*).
void	write_solution_file (const std::string &message) const
	Write the solution file.
void	get_discrete_info (Index &nlvb_, Index &nlvbi_, Index &nlvc_, Index &nlvci_, Index &nlvo_, Index &nlvoi_, Index &nbv_, Index &niv_) const
	ampl orders the variables like (continuous, binary, integer).
Private Member Functions
bool	internal_objval (const Number *x, Number &obj_val)
	Make the objective call to ampl.
bool	internal_conval (const Number x, Index m, Number g=NULL)
	Make the constraint call to ampl.
bool	apply_new_x (bool new_x, Index n, const Number *x)
	Internal function to update the internal and ampl state if the x value changes.
char *	get_options (const SmartPtr< OptionsList > &options, SmartPtr< AmplOptionsList > &ampl_options_list, const char ampl_option_string, const char ampl_invokation_string, const char ampl_banner_string, char *&argv)
	Method for obtaining the name of the NL file and the options set from AMPL.
bool	nerror_ok (void *nerror)
	returns true if the ampl nerror code is ok
void	call_hesset ()
	calls hesset ASL function
Default Compiler Generated Methods
(Hidden to avoid implicit creation/calling). These methods are not implemented and we do not want the compiler to implement them for us, so we declare them private and do not define them. This ensures that they will not be implicitly created/called.
	AmplTNLP ()
	Default Constructor.
	AmplTNLP (const AmplTNLP &)
	Copy Constructor.
void	operator= (const AmplTNLP &)
	Overloaded Equals Operator.
Private Attributes
SmartPtr< const Journalist >	jnlst_
	Journlist.
ASL_pfgh *	asl_
	pointer to the main ASL structure
double	obj_sign_
	Sign of the objective fn (1 for min, -1 for max).
void *	Oinfo_ptr_
	Pointer to the Oinfo structure.
void *	nerror_
	nerror flag passed to ampl calls - set to NULL to halt on error
SmartPtr< AmplSuffixHandler >	suffix_handler_
	Suffix Handler.
Problem Size Data

Index	nz_h_full_
Internal copies of data

Number *	x_sol_
	Solution Vectors.
Number *	z_L_sol_
	Solution Vectors.
Number *	z_U_sol_
	Solution Vectors.
Number *	g_sol_
	Solution Vectors.
Number *	lambda_sol_
	Solution Vectors.
Number	obj_sol_
	Solution Vectors.
Flags to track internal state

bool	objval_called_with_current_x_
	true when the objective value has been calculated with the current x, set to false in apply_new_x, and set to true in internal_objval
bool	conval_called_with_current_x_
	true when the constraint values have been calculated with the current x, set to false in apply_new_x, and set to true in internal_conval
bool	hesset_called_
	true when we have called hesset
bool	set_active_objective_called_
	true when set_active_objective has been called

Detailed Description

Ampl Interface.

Ampl Interface, implemented as a TNLP.

Definition at line 281 of file AmplTNLP.hpp.

Constructor & Destructor Documentation

Ipopt::AmplTNLP::AmplTNLP	(	const SmartPtr< const Journalist > &	jnlst,
		const SmartPtr< OptionsList >	options,
		char **&	argv,
		SmartPtr< AmplSuffixHandler >	suffix_handler = `NULL`,
		bool	allow_discrete = `false`,
		SmartPtr< AmplOptionsList >	ampl_options_list = `NULL`,
		const char *	ampl_option_string = `NULL`,
		const char *	ampl_invokation_string = `NULL`,
		const char *	ampl_banner_string = `NULL`,
		std::string *	nl_file_content = `NULL`
	)

Constructor.

virtual Ipopt::AmplTNLP::~AmplTNLP ( ) [virtual]

Default destructor.

Ipopt::AmplTNLP::AmplTNLP ( ) [private]

Default Constructor.

Ipopt::AmplTNLP::AmplTNLP ( const AmplTNLP & ) [private]

Copy Constructor.

Member Function Documentation

Ipopt::AmplTNLP::DECLARE_STD_EXCEPTION ( NONPOSITIVE_SCALING_FACTOR )

Exceptions.

virtual bool Ipopt::AmplTNLP::get_nlp_info	(	Index &	n,
		Index &	m,
		Index &	nnz_jac_g,
		Index &	nnz_h_lag,
		IndexStyleEnum &	index_style
	)			`[virtual]`

returns dimensions of the nlp.

Overloaded from TNLP

Implements Ipopt::TNLP.

virtual bool Ipopt::AmplTNLP::get_bounds_info	(	Index	n,
		Number *	x_l,
		Number *	x_u,
		Index	m,
		Number *	g_l,
		Number *	g_u
	)			`[virtual]`

returns bounds of the nlp.

Overloaded from TNLP

Implements Ipopt::TNLP.

virtual bool Ipopt::AmplTNLP::get_constraints_linearity	(	Index	m,
		LinearityType *	const_types
	)			`[virtual]`

Returns the constraint linearity.

array will be alocated with length n. (default implementation just return false and does not fill the array).

Reimplemented from Ipopt::TNLP.

virtual bool Ipopt::AmplTNLP::get_starting_point	(	Index	n,
		bool	init_x,
		Number *	x,
		bool	init_z,
		Number *	z_L,
		Number *	z_U,
		Index	m,
		bool	init_lambda,
		Number *	lambda
	)			`[virtual]`

provides a starting point for the nlp variables.

Overloaded from TNLP

Implements Ipopt::TNLP.

virtual bool Ipopt::AmplTNLP::eval_f	(	Index	n,
		const Number *	x,
		bool	new_x,
		Number &	obj_value
	)			`[virtual]`

evaluates the objective value for the nlp.

Overloaded from TNLP

Implements Ipopt::TNLP.

virtual bool Ipopt::AmplTNLP::eval_grad_f	(	Index	n,
		const Number *	x,
		bool	new_x,
		Number *	grad_f
	)			`[virtual]`

evaluates the gradient of the objective for the nlp.

Overloaded from TNLP

Implements Ipopt::TNLP.

virtual bool Ipopt::AmplTNLP::eval_g	(	Index	n,
		const Number *	x,
		bool	new_x,
		Index	m,
		Number *	g
	)			`[virtual]`

evaluates the constraint residuals for the nlp.

Overloaded from TNLP

Implements Ipopt::TNLP.

virtual bool Ipopt::AmplTNLP::eval_jac_g	(	Index	n,
		const Number *	x,
		bool	new_x,
		Index	m,
		Index	nele_jac,
		Index *	iRow,
		Index *	jCol,
		Number *	values
	)			`[virtual]`

specifies the jacobian structure (if values is NULL) and evaluates the jacobian values (if values is not NULL) for the nlp.

Overloaded from TNLP

Implements Ipopt::TNLP.

virtual bool Ipopt::AmplTNLP::eval_h	(	Index	n,
		const Number *	x,
		bool	new_x,
		Number	obj_factor,
		Index	m,
		const Number *	lambda,
		bool	new_lambda,
		Index	nele_hess,
		Index *	iRow,
		Index *	jCol,
		Number *	values
	)			`[virtual]`

specifies the structure of the hessian of the lagrangian (if values is NULL) and evaluates the values (if values is not NULL).

Overloaded from TNLP

Reimplemented from Ipopt::TNLP.

virtual bool Ipopt::AmplTNLP::get_scaling_parameters	(	Number &	obj_scaling,
		bool &	use_x_scaling,
		Index	n,
		Number *	x_scaling,
		bool &	use_g_scaling,
		Index	m,
		Number *	g_scaling
	)			`[virtual]`

retrieve the scaling parameters for the variables, objective function, and constraints.

Reimplemented from Ipopt::TNLP.

virtual void Ipopt::AmplTNLP::finalize_solution	(	SolverReturn	status,
		Index	n,
		const Number *	x,
		const Number *	z_L,
		const Number *	z_U,
		Index	m,
		const Number *	g,
		const Number *	lambda,
		Number	obj_value,
		const IpoptData *	ip_data,
		IpoptCalculatedQuantities *	ip_cq
	)			`[virtual]`

This method is called when the algorithm is complete so the TNLP can store/write the solution.

Implements Ipopt::TNLP.

virtual Index Ipopt::AmplTNLP::get_number_of_nonlinear_variables ( ) [virtual]

Reimplemented from Ipopt::TNLP.

virtual bool Ipopt::AmplTNLP::get_list_of_nonlinear_variables	(	Index	num_nonlin_vars,
		Index *	pos_nonlin_vars
	)			`[virtual]`

Reimplemented from Ipopt::TNLP.

ASL_pfgh* Ipopt::AmplTNLP::AmplSolverObject ( ) [inline]

Return the ampl solver object (ASL*).

Definition at line 386 of file AmplTNLP.hpp.

void Ipopt::AmplTNLP::write_solution_file ( const std::string & message ) const

Write the solution file.

This is a wrapper for AMPL's write_sol. TODO Maybe this should be at a different place, or collect the numbers itself?

void Ipopt::AmplTNLP::get_discrete_info	(	Index &	nlvb_,
		Index &	nlvbi_,
		Index &	nlvc_,
		Index &	nlvci_,
		Index &	nlvo_,
		Index &	nlvoi_,
		Index &	nbv_,
		Index &	niv_
	)			const

ampl orders the variables like (continuous, binary, integer).

This method gives the number of binary and integer variables. For details, see Tables 3 and 4 in "Hooking Your Solver to AMPL"

void Ipopt::AmplTNLP::set_active_objective ( Index obj_no )

A method for setting the index of the objective function to be considered.

This method must be called after the constructor, and before anything else is called. It can only be called once, and if there is more than one objective function in the AMPL model, it MUST be called.

void Ipopt::AmplTNLP::operator= ( const AmplTNLP & ) [private]

Overloaded Equals Operator.

Reimplemented from Ipopt::TNLP.

bool Ipopt::AmplTNLP::internal_objval	(	const Number *	x,
		Number &	obj_val
	)			`[private]`

Make the objective call to ampl.

bool Ipopt::AmplTNLP::internal_conval	(	const Number *	x,
		Index	m,
		Number *	g = `NULL`
	)			`[private]`

Make the constraint call to ampl.

bool Ipopt::AmplTNLP::apply_new_x	(	bool	new_x,
		Index	n,
		const Number *	x
	)			`[private]`

Internal function to update the internal and ampl state if the x value changes.

char* Ipopt::AmplTNLP::get_options	(	const SmartPtr< OptionsList > &	options,
		SmartPtr< AmplOptionsList > &	ampl_options_list,
		const char *	ampl_option_string,
		const char *	ampl_invokation_string,
		const char *	ampl_banner_string,
		char **&	argv
	)			`[private]`