next up previous contents
Next: The OSSolverService Up: Using OSInstance Methods: High Previous: Gradient Evaluation Methods   Contents

Hessian Evaluation Methods

There are two methods for Hessian calculations. The first method has the signature

SparseHessianMatrix *calculateLagrangianHessian( double* x,
    double *objLambda, double *conLambda, bool new_x, int highestOrder);
so if either function or first derivatives have been calculated an appropriate call is
calculateLagrangianHessian( x, w, z, false, 2);
If the Hessian of a single row or objective function is desired the following method is available
SparseHessianMatrix *calculateHessian( double* x, int idx, bool new_x);



Kipp Martin 2008-01-16