Drake
Drake C++ Documentation
non_convex_optimization_util.h File Reference
#include <tuple>
#include <utility>
#include <vector>
#include <Eigen/Core>
#include "drake/solvers/mathematical_program.h"
Include dependency graph for non_convex_optimization_util.h:

drake

drake::solvers

## Functions

std::pair< Eigen::MatrixXd, Eigen::MatrixXd > DecomposeNonConvexQuadraticForm (const Eigen::Ref< const Eigen::MatrixXd > &Q)
For a non-convex homogeneous quadratic form xᵀQx, where Q is not necessarily a positive semidefinite matrix, we decompose it as a difference between two convex homogeneous quadratic forms xᵀQx = xᵀQ₁x - xᵀQ₂x, Q₁, Q₂ are positive semidefinite. More...

std::tuple< Binding< LinearConstraint >, std::vector< Binding< RotatedLorentzConeConstraint > >, VectorXDecisionVariable > AddRelaxNonConvexQuadraticConstraintInTrustRegion (MathematicalProgram *prog, const Eigen::Ref< const VectorXDecisionVariable > &x, const Eigen::Ref< const Eigen::MatrixXd > &Q1, const Eigen::Ref< const Eigen::MatrixXd > &Q2, const Eigen::Ref< const VectorXDecisionVariable > &y, const Eigen::Ref< const Eigen::VectorXd > &p, double lower_bound, double upper_bound, const Eigen::Ref< const Eigen::VectorXd > &linearization_point, double trust_region_gap)
For a non-convex quadratic constraint lb ≤ xᵀQ₁x - xᵀQ₂x + pᵀy ≤ ub where Q₁, Q₂ are both positive semidefinite matrices. More...