This class serves as a base interface for inheriting classes. More...

#include <base_solver_interface.hpp>

Inheritance diagram for DirectTrajectoryOptimization::DTOSolverInterface< DIMENSIONS >:

Public Member Functions
virtual void	InitializeSolver (const std::string &problemName, const std::string &outputFile, const std::string &paramFile, const bool &computeJacobian)=0
	DTO initializes the solver using this method. More...

virtual void	Solve (bool warminit=false)=0
	DTO request the solver to proceed with the NLP. More...

virtual void	GetDTOSolution (state_vector_array_t &yTraj, control_vector_array_t &uTraj, Eigen::VectorXd &hTraj)=0
	Method to get the current values of the solution. More...

virtual void	GetDTOSolution (state_vector_array_t &yTraj, control_vector_array_t &uTraj, Eigen::VectorXd &hTraj, std::vector< int > &phaseId)=0
	Method to get the current values of the solution. More...

virtual void	GetDTOSolution (state_vector_array_t &yTraj, state_vector_array_t &ydotTraj, control_vector_array_t &uTraj, Eigen::VectorXd &hTraj, std::vector< int > &phaseId)
	Method to get the current values of the solution. More...

virtual void	GetSolverSolutionVariables (Eigen::VectorXi &_x_state, Eigen::VectorXd &_x_mul, Eigen::VectorXi &_f_state, Eigen::VectorXd &_f_mul)
	Method to get final values of internal solver variables. More...

virtual void	InitializeDecisionVariablesFromTrajectory (const state_vector_array_t &y_sol, const control_vector_array_t &u_sol, const Eigen::VectorXd &h_sol)=0
	Initial solution point for the solver. More...

virtual void	WarmSolverInitialization (const Eigen::VectorXi &x_state, const Eigen::VectorXd &x_mul, const Eigen::VectorXi &f_state, const Eigen::VectorXd &f_mul)
	Warm Initialization of NLP Solver. More...

virtual void	SetupSolverParameters ()=0
	Setting up solver parameters. More...

virtual void	GetFVector (Eigen::VectorXd &fVec)=0
	Returns the vector of NLP-constraints. More...

void	SetVerbosity (bool verb)
	Set verbosity of the solver. More...

Protected Member Functions
	DTOSolverInterface ()
	Construct an instance of the DTOSolverInterface class. More...

Detailed Description

template<typename DIMENSIONS>
class DirectTrajectoryOptimization::DTOSolverInterface< DIMENSIONS >

This class serves as a base interface for inheriting classes.

Inheriting classes are responsible for implementing interfaces for NLP solvers (eg. IPOPT/SNOPT).

Note: An instance of this class should never be created.

Constructor & Destructor Documentation

template<typename DIMENSIONS >

DirectTrajectoryOptimization::DTOSolverInterface< DIMENSIONS >::DTOSolverInterface ( )

inlineprotected

Construct an instance of the DTOSolverInterface class.

This constructor is not public in order to prevent users from creating this class. This constructor is not private in order to allow derived classes to use the constructor.

Member Function Documentation

template<typename DIMENSIONS >

virtual void DirectTrajectoryOptimization::DTOSolverInterface< DIMENSIONS >::GetDTOSolution	(	state_vector_array_t &	yTraj,
		control_vector_array_t &	uTraj,
		Eigen::VectorXd &	hTraj
	)

pure virtual

Method to get the current values of the solution.

this point is not necessarily feasible or optimal

Parameters

[out]	yTraj	The state space trajectory
[out]	uTraj	The control trajectory
[out]	hTraj	The time increments between the discretization points

Implemented in DirectTrajectoryOptimization::DTOSnoptInterface< DIMENSIONS >, and DirectTrajectoryOptimization::DTOIpoptInterface< DIMENSIONS >.

template<typename DIMENSIONS >

virtual void DirectTrajectoryOptimization::DTOSolverInterface< DIMENSIONS >::GetDTOSolution	(	state_vector_array_t &	yTraj,
		control_vector_array_t &	uTraj,
		Eigen::VectorXd &	hTraj,
		std::vector< int > &	phaseId
	)

pure virtual

Method to get the current values of the solution.

overloaded to also return the phase ID for each node

Parameters

[out]	yTraj	The state space trajectory
[out]	uTraj	The control trajectory
[out]	hTraj	The time increments between the discretization points
[out]	phaseId	The vector of phases ID's of the corresponding nodes

Implemented in DirectTrajectoryOptimization::DTOSnoptInterface< DIMENSIONS >, and DirectTrajectoryOptimization::DTOIpoptInterface< DIMENSIONS >.

template<typename DIMENSIONS >

virtual void DirectTrajectoryOptimization::DTOSolverInterface< DIMENSIONS >::GetDTOSolution	(	state_vector_array_t &	yTraj,
		state_vector_array_t &	ydotTraj,
		control_vector_array_t &	uTraj,
		Eigen::VectorXd &	hTraj,
		std::vector< int > &	phaseId
	)

inlinevirtual

Method to get the current values of the solution.

overloaded to also return the state derivatives

Parameters

[out]	yTraj	The state space trajectory
[out]	ydotTraj	The state derivative trajectory
[out]	uTraj	The control trajectory
[out]	hTraj	The time increments between the discretization points
[out]	phaseId	The vector of phases ID's of the corresponding nodes

Reimplemented in DirectTrajectoryOptimization::DTOSnoptInterface< DIMENSIONS >.

template<typename DIMENSIONS >

virtual void DirectTrajectoryOptimization::DTOSolverInterface< DIMENSIONS >::GetFVector ( Eigen::VectorXd & fVec )

pure virtual

Returns the vector of NLP-constraints.

Parameters

[out] fVec The vector of NLP-constraints

Implemented in DirectTrajectoryOptimization::DTOSnoptInterface< DIMENSIONS >, and DirectTrajectoryOptimization::DTOIpoptInterface< DIMENSIONS >.

template<typename DIMENSIONS >

virtual void DirectTrajectoryOptimization::DTOSolverInterface< DIMENSIONS >::GetSolverSolutionVariables	(	Eigen::VectorXi &	_x_state,
		Eigen::VectorXd &	_x_mul,
		Eigen::VectorXi &	_f_state,
		Eigen::VectorXd &	_f_mul
	)

inlinevirtual

Method to get final values of internal solver variables.

the meaning of the state variables is solver dependent

Parameters

[out]	_x_state	The vector of integers indicating the final state of all decision variables
[out]	_x_mul	The vector of Lagrange multipliers corresponding to all decision variables
[out]	_f_state	The vector of integers indicating the final state of the vector of NLP-constraints
[out]	_f_mul	The vector of Lagrange multipliers corresponding to the vector of NLP-constraints

Reimplemented in DirectTrajectoryOptimization::DTOSnoptInterface< DIMENSIONS >.

template<typename DIMENSIONS >

virtual void DirectTrajectoryOptimization::DTOSolverInterface< DIMENSIONS >::InitializeDecisionVariablesFromTrajectory	(	const state_vector_array_t &	y_sol,
		const control_vector_array_t &	u_sol,
		const Eigen::VectorXd &	h_sol
	)

pure virtual

Initial solution point for the solver.

Set the decision variables to the values given by a trajectory

Parameters

[in]	y_sol	vector of states describing the trajectory
[in]	u_sol	vector of controls describing the trajectory
[in]	h_sol	vector of time increments between discretization points

Implemented in DirectTrajectoryOptimization::DTOSnoptInterface< DIMENSIONS >, and DirectTrajectoryOptimization::DTOIpoptInterface< DIMENSIONS >.

template<typename DIMENSIONS >

virtual void DirectTrajectoryOptimization::DTOSolverInterface< DIMENSIONS >::InitializeSolver	(	const std::string &	problemName,
		const std::string &	outputFile,
		const std::string &	paramFile,
		const bool &	computeJacobian
	)

pure virtual

DTO initializes the solver using this method.

Parameters

[in]	problemName	The name of the problem provided by user
[in]	outputFile	The name of the outputfile provided by user
[in]	paramFile	The name of the parameters file (solver dependent)
[in]	computeJacobian	The flag to use solver numerical differentiation

Implemented in DirectTrajectoryOptimization::DTOSnoptInterface< DIMENSIONS >, and DirectTrajectoryOptimization::DTOIpoptInterface< DIMENSIONS >.

template<typename DIMENSIONS >

virtual void DirectTrajectoryOptimization::DTOSolverInterface< DIMENSIONS >::SetupSolverParameters ( )

pure virtual

Setting up solver parameters.

set solver specific parameters

Implemented in DirectTrajectoryOptimization::DTOSnoptInterface< DIMENSIONS >, and DirectTrajectoryOptimization::DTOIpoptInterface< DIMENSIONS >.

template<typename DIMENSIONS >

void DirectTrajectoryOptimization::DTOSolverInterface< DIMENSIONS >::SetVerbosity ( bool verb )

inline

Set verbosity of the solver.

Parameters

[in] verbose The flag determining the verbosity of the solver

template<typename DIMENSIONS >

virtual void DirectTrajectoryOptimization::DTOSolverInterface< DIMENSIONS >::Solve ( bool warminit = false )

pure virtual

DTO request the solver to proceed with the NLP.

Parameters

[in] warminit The flag indicating the type of optimization

Implemented in DirectTrajectoryOptimization::DTOSnoptInterface< DIMENSIONS >, and DirectTrajectoryOptimization::DTOIpoptInterface< DIMENSIONS >.

template<typename DIMENSIONS >

virtual void DirectTrajectoryOptimization::DTOSolverInterface< DIMENSIONS >::WarmSolverInitialization	(	const Eigen::VectorXi &	x_state,
		const Eigen::VectorXd &	x_mul,
		const Eigen::VectorXi &	f_state,
		const Eigen::VectorXd &	f_mul
	)

inlinevirtual

Warm Initialization of NLP Solver.

set solver variables given by previous solution

Parameters

[in]	x_state	Index for SNOPT decision variables
[in]	x_mul	Lagrange multipliers of decision variables
[in]	f_state	Index for SNOPT vector of constraints
[in]	f_mul	Lagrange multipliers of the constraints

Reimplemented in DirectTrajectoryOptimization::DTOSnoptInterface< DIMENSIONS >.

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

/home/depardo/catkin_ws/src/direct_trajectoryoptimization/include/solver_interfaces/base_solver_interface.hpp

Public Member Functions

Protected Member Functions

Detailed Description

template<typename DIMENSIONS> class DirectTrajectoryOptimization::DTOSolverInterface< DIMENSIONS >

Constructor & Destructor Documentation

Member Function Documentation

template<typename DIMENSIONS>
class DirectTrajectoryOptimization::DTOSolverInterface< DIMENSIONS >