Computing Numerical Derivatives, Gradients and Jacobians
This subroutine computes approximate gradient vectors and Jacobian matrices
of vector functions f(y) in several variables. This routine is used to
compute approximate derivatives for problems where only function
information is available. To approximate a scalar derivative function,
consider it an equivalent 1 x 1 Jacobian matrix. Either one-sided or
central differences can be used.
This software is part of the MATH77/mathc90 collection here.