Module net.finmath.lib

Package net.finmath.montecarlo.interestrate.models.covariance

Interface LIBORCovarianceModel

All Known Subinterfaces:

LIBORCovarianceModelCalibrateable

All Known Implementing Classes:

AbstractLIBORCovarianceModel, AbstractLIBORCovarianceModelParametric, BlendedLocalVolatilityModel, DisplacedLocalVolatilityModel, ExponentialDecayLocalVolatilityModel, HullWhiteLocalVolatilityModel, LIBORCovarianceModelBH, LIBORCovarianceModelExponentialForm5Param, LIBORCovarianceModelExponentialForm7Param, LIBORCovarianceModelFromVolatilityAndCorrelation, LIBORCovarianceModelStochasticHestonVolatility, LIBORCovarianceModelStochasticVolatility

public interface LIBORCovarianceModel

Interface for covariance models providing a vector of (possibly stochastic) factor loadings. Classes implementing this interface can be used as "plug ins" for LIBORMarketModelFromCovarianceModel.

Version:

1.0

Author:

Christian Fries

Method Summary

Modifier and Type	Method	Description
AbstractLIBORCovarianceModelParametric	getCloneWithModifiedData(Map<String,Object> dataModified)	Returns a clone of this model where the specified properties have been modified.
RandomVariable	getCovariance(double time, int component1, int component2, RandomVariable[] realizationAtTimeIndex)	Returns the instantaneous covariance calculated from factor loadings.
RandomVariable	getCovariance(int timeIndex, int component1, int component2, RandomVariable[] realizationAtTimeIndex)	Returns the instantaneous covariance calculated from factor loadings.
RandomVariable[]	getFactorLoading(double time, double component, RandomVariable[] realizationAtTimeIndex)	Return the factor loading for a given time and a given component.
RandomVariable[]	getFactorLoading(double time, int component, RandomVariable[] realizationAtTimeIndex)	Return the factor loading for a given time and component index.
RandomVariable[]	getFactorLoading(int timeIndex, int component, RandomVariable[] realizationAtTimeIndex)	Return the factor loading for a given time index and component index.
RandomVariable	getFactorLoadingPseudoInverse(int timeIndex, int component, int factor, RandomVariable[] realizationAtTimeIndex)	Returns the pseudo inverse of the factor matrix.
TimeDiscretization	getLiborPeriodDiscretization()	The forward rate time discretization associated with this model (defines the components).
int	getNumberOfFactors()
TimeDiscretization	getTimeDiscretization()	The simulation time discretization associated with this model.

Method Detail

getFactorLoading

RandomVariable[] getFactorLoading(double time,
                                  double component,
                                  RandomVariable[] realizationAtTimeIndex)

Return the factor loading for a given time and a given component. The factor loading is the vector f_i such that the scalar product
f_jf_k = f_j,1f_k,1 + ... + f_j,mf_k,m
is the instantaneous covariance of the component j and k. With respect to simulation time t, this method uses a piece wise constant interpolation, i.e., it calculates t__i such that t__i is the largest point in getTimeDiscretization such that t__i ≤ t . The component here, it given via a double T which may be associated with the LIBOR fixing date. With respect to component time T, this method uses a piece wise constant interpolation, i.e., it calculates T__j such that T__j is the largest point in getTimeDiscretization such that T__j ≤ T .

Parameters:

time - The time t at which factor loading is requested.

component - The component time (as a double associated with the fixing of the forward rate) T_i.

realizationAtTimeIndex - The realization of the stochastic process (may be used to implement local volatility/covariance/correlation models).

Returns:

The factor loading f_i(t).

getFactorLoading

RandomVariable[] getFactorLoading(double time,
                                  int component,
                                  RandomVariable[] realizationAtTimeIndex)

Return the factor loading for a given time and component index. The factor loading is the vector f_i such that the scalar product
f_jf_k = f_j,1f_k,1 + ... + f_j,mf_k,m
is the instantaneous covariance of the component j and k. With respect to simulation time t, this method uses a piece wise constant interpolation, i.e., it calculates t__i such that t__i is the largest point in getTimeDiscretization such that t__i ≤ t .

Parameters:

time - The time t at which factor loading is requested.

component - The index of the component i. Note that this class may have its own LIBOR time discretization and that this index refers to this discretization.

realizationAtTimeIndex - The realization of the stochastic process (may be used to implement local volatility/covariance/correlation models).

Returns:

The factor loading f_i(t).

getFactorLoading

RandomVariable[] getFactorLoading(int timeIndex,
                                  int component,
                                  RandomVariable[] realizationAtTimeIndex)

Return the factor loading for a given time index and component index. The factor loading is the vector f_i such that the scalar product
f_jf_k = f_j,1f_k,1 + ... + f_j,mf_k,m
is the instantaneous covariance of the component j and k.

Parameters:

timeIndex - The time index at which factor loading is requested.

component - The index of the component i.

realizationAtTimeIndex - The realization of the stochastic process (may be used to implement local volatility/covariance/correlation models).

Returns:

The factor loading f_i(t).

getFactorLoadingPseudoInverse

RandomVariable getFactorLoadingPseudoInverse(int timeIndex,
                                             int component,
                                             int factor,
                                             RandomVariable[] realizationAtTimeIndex)

Returns the pseudo inverse of the factor matrix.

Parameters:

timeIndex - The time index at which factor loading inverse is requested.

factor - The index of the factor j.

component - The index of the component i.

realizationAtTimeIndex - The realization of the stochastic process (may be used to implement local volatility/covariance/correlation models).

Returns:

The entry of the pseudo-inverse of the factor loading matrix.

getCovariance

RandomVariable getCovariance(double time,
                             int component1,
                             int component2,
                             RandomVariable[] realizationAtTimeIndex)

Returns the instantaneous covariance calculated from factor loadings.

Parameters:

time - The time t at which covariance is requested.

component1 - Index of component i.

component2 - Index of component j.

realizationAtTimeIndex - The realization of the stochastic process.

Returns:

The instantaneous covariance between component i and j.

getCovariance

RandomVariable getCovariance(int timeIndex,
                             int component1,
                             int component2,
                             RandomVariable[] realizationAtTimeIndex)

Returns the instantaneous covariance calculated from factor loadings.

Parameters:

timeIndex - The time index at which covariance is requested.

component1 - Index of component i.

component2 - Index of component j.

realizationAtTimeIndex - The realization of the stochastic process.

Returns:

The instantaneous covariance between component i and j.

getTimeDiscretization

TimeDiscretization getTimeDiscretization()

The simulation time discretization associated with this model.

Returns:

the timeDiscretizationFromArray

getLiborPeriodDiscretization

TimeDiscretization getLiborPeriodDiscretization()

The forward rate time discretization associated with this model (defines the components).

Returns:

the forward rate time discretization associated with this model.

getNumberOfFactors

int getNumberOfFactors()

Returns:

the numberOfFactors

getCloneWithModifiedData

AbstractLIBORCovarianceModelParametric getCloneWithModifiedData(Map<String,Object> dataModified)
                                                         throws CalculationException

Returns a clone of this model where the specified properties have been modified. Note that there is no guarantee that a model reacts on a specification of a properties in the parameter map dataModified. If data is provided which is ignored by the model no exception may be thrown. Furthermore the structure of the covariance model has to match changed data. A change of the time discretizations may requires a change in the parameters but this function will just insert the new time discretization without changing the parameters. An exception may not be thrown.

Parameters:

dataModified - Key-value-map of parameters to modify.

Returns:

A clone of this model (or a new instance of this model if no parameter was modified).

Throws:

CalculationException - Thrown when the model could not be created.