|
ROL
|
Provides an interface for the risk measure associated with the super quantile quadrangle. More...
#include <ROL_SecondOrderCVaR.hpp>
Public Member Functions | |
| SecondOrderCVaR (ROL::ParameterList &parlist) | |
| SecondOrderCVaR (const Real alpha, const int nQuad, const ROL::Ptr< PlusFunction< Real > > &pf, const bool useGauss=true) | |
| Public Member Functions inherited from ROL::SpectralRisk< Real > | |
| SpectralRisk (void) | |
| SpectralRisk (const ROL::Ptr< Distribution< Real > > &dist, const int nQuad, const ROL::Ptr< PlusFunction< Real > > &pf) | |
| SpectralRisk (ROL::ParameterList &parlist) | |
| SpectralRisk (const std::vector< Real > &pts, const std::vector< Real > &wts, const ROL::Ptr< PlusFunction< Real > > &pf) | |
| void | setStorage (const Ptr< ScalarController< Real > > &value_storage, const Ptr< VectorController< Real > > &gradient_storage) override |
| void | setHessVecStorage (const Ptr< ScalarController< Real > > &gradvec_storage, const Ptr< VectorController< Real > > &hessvec_storage) override |
| void | setSample (const std::vector< Real > &point, const Real weight) override |
| void | resetStorage (bool flag=true) override |
| void | resetStorage (UpdateType type) override |
| void | initialize (const Vector< Real > &x) override |
| Real | computeStatistic (const Ptr< const std::vector< Real > > &xstat) const override |
| void | updateValue (Objective< Real > &obj, const Vector< Real > &x, const std::vector< Real > &xstat, Real &tol) override |
| void | updateGradient (Objective< Real > &obj, const Vector< Real > &x, const std::vector< Real > &xstat, Real &tol) override |
| void | updateHessVec (Objective< Real > &obj, const Vector< Real > &v, const std::vector< Real > &vstat, const Vector< Real > &x, const std::vector< Real > &xstat, Real &tol) override |
| Real | getValue (const Vector< Real > &x, const std::vector< Real > &xstat, SampleGenerator< Real > &sampler) override |
| void | getGradient (Vector< Real > &g, std::vector< Real > &gstat, const Vector< Real > &x, const std::vector< Real > &xstat, SampleGenerator< Real > &sampler) override |
| void | getHessVec (Vector< Real > &hv, std::vector< Real > &hvstat, const Vector< Real > &v, const std::vector< Real > &vstat, const Vector< Real > &x, const std::vector< Real > &xstat, SampleGenerator< Real > &sampler) override |
Private Member Functions | |
| void | checkInputs (void) const |
| void | initializeQuad (void) |
Private Attributes | |
| ROL::Ptr< PlusFunction< Real > > | plusFunction_ |
| Real | alpha_ |
| int | nQuad_ |
| bool | useGauss_ |
| std::vector< Real > | wts_ |
| std::vector< Real > | pts_ |
Additional Inherited Members | |
| Protected Member Functions inherited from ROL::SpectralRisk< Real > | |
| void | buildMixedQuantile (const std::vector< Real > &pts, const std::vector< Real > &wts, const ROL::Ptr< PlusFunction< Real > > &pf) |
| void | buildQuadFromDist (std::vector< Real > &pts, std::vector< Real > &wts, const int nQuad, const ROL::Ptr< Distribution< Real > > &dist) const |
| void | printQuad (const std::vector< Real > &pts, const std::vector< Real > &wts, const bool print=false) const |
Provides an interface for the risk measure associated with the super quantile quadrangle.
The risk measure associated with the super quantile quadrangle is defined as
\[ \mathcal{R}(X) = \frac{1}{1-\beta}\int_\beta^1\mathrm{CVaR}_{\alpha}(X) \,\mathrm{d}\alpha \]
where \(0 \le \beta < 1\) and the conditional value-at-risk (CVaR) with confidence level \(0\le \alpha < 1\) is
\[ \mathrm{CVaR}_\alpha(X) = \inf_{t\in\mathbb{R}} \left\{ t + \frac{1}{1-\alpha} \mathbb{E}\left[(X-t)_+\right] \right\} \]
where \((x)_+ = \max\{0,x\}\). If the distribution of \(X\) is continuous, then \(\mathrm{CVaR}_{\alpha}(X)\) is the conditional expectation of \(X\) exceeding the \(\alpha\)-quantile of \(X\) and the optimal \(t\) is the \(\alpha\)-quantile. Additionally, \(\mathcal{R}\) is a law-invariant coherent risk measure.
ROL implements \(\mathcal{R}\) by approximating the integral with Gauss-Legendre or Fejer 2 quadrature. The corresponding quadrature points and weights are then used to construct a ROL::MixedQuantileQuadrangle risk measure. When using derivative-based optimization, the user can provide a smooth approximation of \((\cdot)_+\) using the ROL::PlusFunction class.
Definition at line 51 of file ROL_SecondOrderCVaR.hpp.
|
inline |
Definition at line 92 of file ROL_SecondOrderCVaR.hpp.
References alpha_, checkInputs(), initializeQuad(), nQuad_, plusFunction_, ROL::SpectralRisk< Real >::SpectralRisk(), and useGauss_.
|
inline |
Definition at line 106 of file ROL_SecondOrderCVaR.hpp.
References alpha_, checkInputs(), initializeQuad(), nQuad_, plusFunction_, ROL::SpectralRisk< Real >::SpectralRisk(), and useGauss_.
|
inlineprivate |
Definition at line 62 of file ROL_SecondOrderCVaR.hpp.
References alpha_, and plusFunction_.
Referenced by SecondOrderCVaR(), and SecondOrderCVaR().
|
inlineprivate |
Definition at line 69 of file ROL_SecondOrderCVaR.hpp.
References alpha_, ROL::SpectralRisk< Real >::buildMixedQuantile(), nQuad_, plusFunction_, pts_, useGauss_, and wts_.
Referenced by SecondOrderCVaR(), and SecondOrderCVaR().
|
private |
Definition at line 53 of file ROL_SecondOrderCVaR.hpp.
Referenced by checkInputs(), initializeQuad(), SecondOrderCVaR(), and SecondOrderCVaR().
|
private |
Definition at line 55 of file ROL_SecondOrderCVaR.hpp.
Referenced by checkInputs(), initializeQuad(), SecondOrderCVaR(), and SecondOrderCVaR().
|
private |
Definition at line 56 of file ROL_SecondOrderCVaR.hpp.
Referenced by initializeQuad(), SecondOrderCVaR(), and SecondOrderCVaR().
|
private |
Definition at line 57 of file ROL_SecondOrderCVaR.hpp.
Referenced by initializeQuad(), SecondOrderCVaR(), and SecondOrderCVaR().
|
private |
Definition at line 59 of file ROL_SecondOrderCVaR.hpp.
Referenced by initializeQuad().
|
private |
Definition at line 60 of file ROL_SecondOrderCVaR.hpp.
Referenced by initializeQuad().