Classes | |
| struct | PrimalDualDerivativeResult |
| This class stores the results of the primal-dual line-search. More... | |
Functions | |
| template<typename T> | |
| auto | gpdal_derivative_results (const Model< T > &qpmodel, Results< T > &qpresults, Workspace< T > &qpwork, const Settings< T > &qpsettings, isize n_constraints, T alpha) -> PrimalDualDerivativeResult< T > |
| template<typename T> | |
| auto | primal_dual_derivative_results (const Model< T > &qpmodel, Results< T > &qpresults, Workspace< T > &qpwork, isize n_constraints, T alpha) -> PrimalDualDerivativeResult< T > |
| template<typename T> | |
| void | primal_dual_ls (const Model< T > &qpmodel, Results< T > &qpresults, Workspace< T > &qpwork, const Settings< T > &qpsettings, const isize n_constraints) |
| template<typename T> | |
| void | active_set_change (const Model< T > &qpmodel, Results< T > &qpresults, const DenseBackend &dense_backend, const isize n_constraints, Workspace< T > &qpwork) |
| auto proxsuite::proxqp::dense::linesearch::gpdal_derivative_results | ( | const Model< T > & | qpmodel, |
| Results< T > & | qpresults, | ||
| Workspace< T > & | qpwork, | ||
| const Settings< T > & | qpsettings, | ||
| isize | n_constraints, | ||
| T | alpha ) -> PrimalDualDerivativeResult<T> |
Stores first derivative and coefficient of the univariate second order polynomial merit function to be canceled by the exact generalized primal-dual linesearch.
| a | second order polynomial coefficient of the merit function used in the linesearch. |
| b | first order polynomial coefficient of the merit function used in the linesearch. |
| grad | derivative of the merit function used in the linesearch. |
Definition at line 51 of file linesearch.hpp.
| auto proxsuite::proxqp::dense::linesearch::primal_dual_derivative_results | ( | const Model< T > & | qpmodel, |
| Results< T > & | qpresults, | ||
| Workspace< T > & | qpwork, | ||
| isize | n_constraints, | ||
| T | alpha ) -> PrimalDualDerivativeResult<T> |
Stores first derivative and coefficient of the univariate second order polynomial merit function to be canceled by the exact primal-dual linesearch.
| a | second order polynomial coefficient of the merit function used in the linesearch. |
| b | first order polynomial coefficient of the merit function used in the linesearch. |
| grad | derivative of the merit function used in the linesearch. |
Definition at line 180 of file linesearch.hpp.
| void proxsuite::proxqp::dense::linesearch::primal_dual_ls | ( | const Model< T > & | qpmodel, |
| Results< T > & | qpresults, | ||
| Workspace< T > & | qpwork, | ||
| const Settings< T > & | qpsettings, | ||
| const isize | n_constraints ) |
Performs the exact primaldual linesearch algorithm.
| qpwork | solver workspace. |
| qpmodel | QP problem model as defined by the user (without any scaling performed). |
| qpresults | solver results. |
Definition at line 322 of file linesearch.hpp.
| void proxsuite::proxqp::dense::linesearch::active_set_change | ( | const Model< T > & | qpmodel, |
| Results< T > & | qpresults, | ||
| const DenseBackend & | dense_backend, | ||
| const isize | n_constraints, | ||
| Workspace< T > & | qpwork ) |
Performs the active set change of the factorized KKT matrix (using rank one updates or downgrades).
| qpwork | solver workspace. |
| qpmodel | QP problem model as defined by the user (without any scaling performed). |
| qpresults | solver results. |
Definition at line 551 of file linesearch.hpp.