Quadratic Program solver. More...

#include <chrono>
#include <iomanip>
#include <iostream>
#include <limits>
#include <optional>
#include <Eigen/Cholesky>
#include <Eigen/Dense>
#include <Eigen/Sparse>
#include <Eigen/SparseCholesky>
#include "qp.hpp"
#include "utils/sparse.hpp"

Include dependency graph for qp_solver.hpp:

This graph shows which files directly or indirectly include this file:

Go to the source code of this file.

Classes
struct	smooth::feedback::QPSolverParams
	Options for solve_qp. More...

struct	smooth::feedback::detail::LDLTWrapper< LDLTt >
	Copy-able wrapper around Eigen LDLT types. More...

class	smooth::feedback::QPSolver< Pbm >
	Solver for quadratic programs. More...

Typedefs
template<typename Pbm >
using	smooth::feedback::detail::qp_solution_t = QPSolution< decltype(Pbm::A)::RowsAtCompileTime, decltype(Pbm::A)::ColsAtCompileTime, typename decltype(Pbm::A)::Scalar >

Functions
template<typename Pbm , typename D1 , typename D2 >
bool	smooth::feedback::detail::polish_qp (const Pbm &pbm, qp_solution_t< Pbm > &sol, const QPSolverParams &prm, const typename decltype(Pbm::A)::Scalar c, const Eigen::MatrixBase< D1 > &sx, const Eigen::MatrixBase< D2 > &sy)
	Polish solution of quadratic program. More...

template<typename Pbm >
detail::qp_solution_t< Pbm >	smooth::feedback::solve_qp (const Pbm &pbm, const QPSolverParams &prm, std::optional< std::reference_wrapper< const detail::qp_solution_t< Pbm > > > warmstart={})
	Solve a quadratic program using the operator splitting approach. More...

Detailed Description

Quadratic Program solver.

Definition in file qp_solver.hpp.

Typedef Documentation

◆ qp_solution_t

template<typename Pbm >

using smooth::feedback::detail::qp_solution_t = typedef QPSolution< decltype(Pbm::A)::RowsAtCompileTime, decltype(Pbm::A)::ColsAtCompileTime, typename decltype(Pbm::A)::Scalar>

Definition at line 73 of file qp_solver.hpp.

Function Documentation

◆ polish_qp()

template<typename Pbm , typename D1 , typename D2 >

bool smooth::feedback::detail::polish_qp	(	const Pbm &	pbm,
		qp_solution_t< Pbm > &	sol,
		const QPSolverParams &	prm,
		const typename decltype(Pbm::A)::Scalar	c,
		const Eigen::MatrixBase< D1 > &	sx,
		const Eigen::MatrixBase< D2 > &	sy
	)

Polish solution of quadratic program.

Template Parameters

Pbm	problem type

Parameters

[in]	pbm	problem formulation
[in,out]	sol	solution to polish
[in]	prm	solver options
[in]	c	cost scaling
[in]	sx	variable scaling
[in]	sy	constraint scaling

Warning: This function allocates dynamic memory even for static-sized problems.

Definition at line 93 of file qp_solver.hpp.

◆ solve_qp()

template<typename Pbm >

detail::qp_solution_t< Pbm > smooth::feedback::solve_qp	(	const Pbm &	pbm,
		const QPSolverParams &	prm,
		std::optional< std::reference_wrapper< const detail::qp_solution_t< Pbm > > >	warmstart = `{}`
	)

Solve a quadratic program using the operator splitting approach.

Template Parameters

Pbm	problem type (QuadraticProgram or QuadraticProgramSparse)

Parameters

pbm	problem formulation
prm	solver options
warmstart	provide initial guess for primal and dual variables

Returns: solution as QuasraticProgramSolution<M, N>

Note: dynamic problem sizes (M == -1 || N == -1) are supported

This is a third-party implementation of the algorithm described in the following paper:

Stellato, B., Banjac, G., Goulart, P. et al. OSQP: an operator splitting solver for quadratic programs. Math. Prog. Comp. 12, 637–672 (2020). https://doi.org/10.1007/s12532-020-00179-2

For the official C implementation, see https://osqp.org/.

Definition at line 780 of file qp_solver.hpp.

Classes