CppAD: A C++ Algorithmic Differentiation Package
20130102
|
size_t ADFun< Base >::SparseJacobianReverse | ( | const VectorBase & | x, |
const VectorSet & | p, | ||
const VectorSize & | row, | ||
const VectorSize & | col, | ||
VectorBase & | jac, | ||
sparse_jacobian_work & | work | ||
) |
Compute user specified subset of a sparse Jacobian using forward mode.
The C++ source code corresponding to this operation is
SparceJacobianReverse(x, p, row, col, jac, work)
Base | See SparseJacobianForward(x, p, row, col, jac, work) . |
VectorBase | See SparseJacobianForward(x, p, row, col, jac, work) . |
VectorSet | See SparseJacobianForward(x, p, row, col, jac, work) . |
VectorSize | See SparseJacobianForward(x, p, row, col, jac, work) . |
x | See SparseJacobianForward(x, p, row, col, jac, work) . |
p | See SparseJacobianForward(x, p, row, col, jac, work) . |
row | See SparseJacobianForward(x, p, row, col, jac, work) . |
col | See SparseJacobianForward(x, p, row, col, jac, work) . |
jac | See SparseJacobianForward(x, p, row, col, jac, work) . |
work | work contains information that depends on the function object, sparsity pattern, row vector, and col vector. These work values are computed by SparseJacobianReverse . If they are not the same between calls, work.clear() must be called to reinitialize them. To be more specific: user_row is a copy of row with the value m at the end (m is the range dimension), user_col is a copy of col with the value n at the end (n is the domain dimension), sort_row is the index vector that sorts the user_row vector with the extra value K at the end (K is the size of jac ). |
Definition at line 946 of file sparse_jacobian.hpp.