template<class Base >

template<class SizeVector , class BaseVector >

size_t CppAD::ADFun< Base >::sparse_hes	(	const BaseVector &	x,
		const BaseVector &	w,
		sparse_rcv< SizeVector, BaseVector > &	subset,
		const sparse_rc< SizeVector > &	pattern,
		const std::string &	coloring,
		sparse_hes_work &	work
	)

Calculate sparse Hessians using forward mode.

Template Parameters

Base	the base type for the recording that is stored in the ADFun object.
SizeVector	a simple vector class with elements of type size_t.
BaseVector	a simple vector class with elements of type Base.

Parameters

x	a vector of length n, the number of independent variables in f (this ADFun object).
w	a vector of length m, the number of dependent variables in f (this ADFun object).
subset	specifices the subset of the sparsity pattern where the Hessian is evaluated. subset.nr() == n, subset.nc() == n.
pattern	is a sparsity pattern for the Hessian of w^T * f; pattern.nr() == n, pattern.nc() == n, where m is number of dependent variables in f.
coloring	determines which coloring algorithm is used. This must be cppad.symmetric, cppad.general, colpack.symmetic, or colpack.star.
work	this structure must be empty, or contain the information stored by a previous call to sparse_hes. The previous call must be for the same ADFun object f and the same subset.

Returns

This is the number of first order forward (and second order reverse) sweeps used to compute thhe Hessian.

Definition at line 318 of file sparse_hes.hpp.