numerical cost for path optimization More...
#include <hpp/core/path-optimization/cost.hh>
Public Member Functions | |
virtual void | hessian (matrixOut_t hessian) const =0 |
Return an approximation of the Hessian at minimum. More... | |
![]() | |
virtual | ~DifferentiableFunction () |
LiegroupElement | operator() (vectorIn_t argument) const |
void | value (LiegroupElement &result, vectorIn_t argument) const |
void | jacobian (matrixOut_t jacobian, vectorIn_t argument) const |
const ArrayXb & | activeParameters () const |
const ArrayXb & | activeDerivativeParameters () const |
size_type | inputSize () const |
size_type | inputDerivativeSize () const |
LiegroupSpacePtr_t | outputSpace () const |
size_type | outputSize () const |
size_type | outputDerivativeSize () const |
const std::string & | name () const |
virtual std::ostream & | print (std::ostream &o) const |
std::string | context () const |
void | context (const std::string &c) |
void | finiteDifferenceForward (matrixOut_t jacobian, vectorIn_t arg, DevicePtr_t robot=DevicePtr_t(), value_type eps=std::sqrt(Eigen::NumTraits< value_type >::epsilon())) const |
void | finiteDifferenceCentral (matrixOut_t jacobian, vectorIn_t arg, DevicePtr_t robot=DevicePtr_t(), value_type eps=std::sqrt(Eigen::NumTraits< value_type >::epsilon())) const |
Protected Member Functions | |
Cost (size_type inputSize, size_type inputDerivativeSize, const std::string &name) | |
![]() | |
DifferentiableFunction (size_type sizeInput, size_type sizeInputDerivative, size_type sizeOutput, std::string name=std::string()) | |
DifferentiableFunction (size_type sizeInput, size_type sizeInputDerivative, const LiegroupSpacePtr_t &outputSpace, std::string name=std::string()) | |
virtual void | impl_compute (LiegroupElement &result, vectorIn_t argument) const=0 |
virtual void | impl_jacobian (matrixOut_t jacobian, vectorIn_t arg) const=0 |
Additional Inherited Members | |
![]() | |
size_type | inputSize_ |
size_type | inputDerivativeSize_ |
LiegroupSpacePtr_t | outputSpace_ |
ArrayXb | activeParameters_ |
ArrayXb | activeDerivativeParameters_ |
numerical cost for path optimization
Provides an initial guess for the Hessian to initialize quasi-Newton methods.
|
inlineprotected |
|
pure virtual |
Return an approximation of the Hessian at minimum.
hessian | Hessian matrix of right size |
Implemented in hpp::core::pathOptimization::PathLength.