A C++ example for interfacing an MINLP with bonmin. More...

#include <MyTMINLP.hpp>

Inheritance diagram for MyTMINLP:

Public Member Functions
	MyTMINLP ()
	Default constructor.
virtual	~MyTMINLP ()
	virtual destructor.
	MyTMINLP (const MyTMINLP &other)
	Copy constructor.
virtual const SosInfo *	sosConstraints () const
virtual const BranchingInfo *	branchingInfo () const
void	printSolutionAtEndOfAlgorithm ()
Overloaded functions specific to a TMINLP.
Assignment operator. no data = nothing to assign
virtual bool	get_variables_types (Index n, VariableType *var_types)
	Pass the type of the variables (INTEGER, BINARY, CONTINUOUS) to the optimizer.
virtual bool	get_variables_linearity (Index n, Ipopt::TNLP::LinearityType *var_types)
	Pass info about linear and nonlinear variables.
virtual bool	get_constraints_linearity (Index m, Ipopt::TNLP::LinearityType *const_types)
	Pass the type of the constraints (LINEAR, NON_LINEAR) to the optimizer.
Overloaded functions defining a TNLP.
This group of function implement the various elements needed to define and solve a TNLP. They are the same as those in a standard Ipopt NLP problem
virtual bool	get_nlp_info (Index &n, Index &m, Index &nnz_jac_g, Index &nnz_h_lag, TNLP::IndexStyleEnum &index_style)
	Method to pass the main dimensions of the problem to Ipopt.
virtual bool	get_bounds_info (Index n, Number x_l, Number x_u, Index m, Number g_l, Number g_u)
	Method to pass the bounds on variables and constraints to Ipopt.
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)
	Method to to pass the starting point for optimization to Ipopt.
virtual bool	eval_f (Index n, const Number *x, bool new_x, Number &obj_value)
	Method which compute the value of the objective function at point x.
virtual bool	eval_grad_f (Index n, const Number x, bool new_x, Number grad_f)
	Method which compute the gradient of the objective at a point x.
virtual bool	eval_g (Index n, const Number x, bool new_x, Index m, Number g)
	Method which compute the value of the functions defining the constraints at a point x.
virtual bool	eval_jac_g (Index n, const Number x, bool new_x, Index m, Index nele_jac, Index iRow, Index jCol, Number values)
	Method to compute the Jacobian of the functions defining the constraints.
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)
	Method to compute the Jacobian of the functions defining the constraints.
virtual void	finalize_solution (TMINLP::SolverReturn status, Index n, const Number *x, Number obj_value)
	Method called by Ipopt at the end of optimization.
Private Attributes
bool	printSol_

Detailed Description

A C++ example for interfacing an MINLP with bonmin.

This class implements the following NLP :

$\begin{array}{l} \min - x_0 - x_1 - x_2 \\ \mbox{s.t}\\ (x_1 - \frac{1}{2})^2 + (x_2 - \frac{1}{2})^2 \leq \frac{1}{4} \\ x_0 - x_1 \leq 0 \\ x_0 + x_2 + x_3 \leq 2\\ x_0 \in \{0,1\}^n \; (x_1, x_2) \in R^2 \; x_3 \in N \end{array}$

Definition at line 28 of file MyTMINLP.hpp.

Constructor & Destructor Documentation

MyTMINLP::MyTMINLP ( ) [inline]

Default constructor.

Definition at line 32 of file MyTMINLP.hpp.

virtual MyTMINLP::~MyTMINLP ( ) [inline, virtual]

virtual destructor.

Definition at line 36 of file MyTMINLP.hpp.

MyTMINLP::MyTMINLP ( const MyTMINLP & other ) [inline]

Copy constructor.

Definition at line 40 of file MyTMINLP.hpp.

Member Function Documentation

bool MyTMINLP::get_variables_types	(	Index	n,
		VariableType *	var_types
	)		`[virtual]`

Pass the type of the variables (INTEGER, BINARY, CONTINUOUS) to the optimizer.

Parameters:

n	size of var_types (has to be equal to the number of variables in the problem)
var_types	types of the variables (has to be filled by function).

Definition at line 13 of file MyTMINLP.cpp.

References Bonmin::TMINLP::BINARY, Bonmin::TMINLP::CONTINUOUS, and Bonmin::TMINLP::INTEGER.

bool MyTMINLP::get_variables_linearity	(	Index	n,
		Ipopt::TNLP::LinearityType *	var_types
	)		`[virtual]`

Pass info about linear and nonlinear variables.

Definition at line 24 of file MyTMINLP.cpp.

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

Pass the type of the constraints (LINEAR, NON_LINEAR) to the optimizer.

Parameters:

m	size of const_types (has to be equal to the number of constraints in the problem)
const_types	types of the constraints (has to be filled by function).

Definition at line 35 of file MyTMINLP.cpp.

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

Method to pass the main dimensions of the problem to Ipopt.

Parameters:

n	number of variables in problem.
m	number of constraints.
nnz_jac_g	number of nonzeroes in Jacobian of constraints system.
nnz_h_lag	number of nonzeroes in Hessian of the Lagrangean.
index_style	indicate wether arrays are numbered from 0 (C-style) or from 1 (Fortran).

Returns:: true in case of success.

Definition at line 44 of file MyTMINLP.cpp.

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

Method to pass the bounds on variables and constraints to Ipopt.

Parameters:

n	size of x_l and x_u (has to be equal to the number of variables in the problem)
x_l	lower bounds on variables (function should fill it).
x_u	upper bounds on the variables (function should fill it).
m	size of g_l and g_u (has to be equal to the number of constraints in the problem).
g_l	lower bounds of the constraints (function should fill it).
g_u	upper bounds of the constraints (function should fill it).

Returns:: true in case of success.

Definition at line 56 of file MyTMINLP.cpp.

bool MyTMINLP::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]`

Method to to pass the starting point for optimization to Ipopt.

Parameters:

init_x	do we initialize primals?
x	pass starting primal points (function should fill it if init_x is 1).
m	size of lambda (has to be equal to the number of constraints in the problem).
init_lambda	do we initialize duals of constraints?
lambda	lower bounds of the constraints (function should fill it).

Returns:: true in case of success.

Definition at line 85 of file MyTMINLP.cpp.

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

Method which compute the value of the objective function at point x.

Parameters:

n	size of array x (has to be the number of variables in the problem).
x	point where to evaluate.
new_x	Is this the first time we evaluate functions at this point? (in the present context we don't care).
obj_value	value of objective in x (has to be computed by the function).

Returns:: true in case of success.

Definition at line 103 of file MyTMINLP.cpp.

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

Method which compute the gradient of the objective at a point x.

Parameters:

n	size of array x (has to be the number of variables in the problem).
x	point where to evaluate.
new_x	Is this the first time we evaluate functions at this point? (in the present context we don't care).
grad_f	gradient of objective taken in x (function has to fill it).

Returns:: true in case of success.

Definition at line 111 of file MyTMINLP.cpp.

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

Method which compute the value of the functions defining the constraints at a point x.

Parameters:

n	size of array x (has to be the number of variables in the problem).
x	point where to evaluate.
new_x	Is this the first time we evaluate functions at this point? (in the present context we don't care).
m	size of array g (has to be equal to the number of constraints in the problem)
grad_f	values of the constraints (function has to fill it).

Returns:: true in case of success.

Definition at line 122 of file MyTMINLP.cpp.

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

Method to compute the Jacobian of the functions defining the constraints.

If the parameter values==NULL fill the arrays iCol and jRow which store the position of the non-zero element of the Jacobian. If the paramenter values!=NULL fill values with the non-zero elements of the Jacobian.

Parameters:

n	size of array x (has to be the number of variables in the problem).
x	point where to evaluate.
new_x	Is this the first time we evaluate functions at this point? (in the present context we don't care).
m	size of array g (has to be equal to the number of constraints in the problem)
grad_f	values of the constraints (function has to fill it).

Returns:: true in case of success.

Definition at line 135 of file MyTMINLP.cpp.

bool MyTMINLP::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]`

Method to compute the Jacobian of the functions defining the constraints.

If the parameter values==NULL fill the arrays iCol and jRow which store the position of the non-zero element of the Jacobian. If the paramenter values!=NULL fill values with the non-zero elements of the Jacobian.

Parameters:

n	size of array x (has to be the number of variables in the problem).
x	point where to evaluate.
new_x	Is this the first time we evaluate functions at this point? (in the present context we don't care).
m	size of array g (has to be equal to the number of constraints in the problem)
grad_f	values of the constraints (function has to fill it).