Utility of Jacobian in lsqnonlin

Hi,

I have a system of nonlinear equations and I solved it with lsqnonlin with and without the jacobian.

The only difference I saw is the computation time that is much smaller with the jacobian ( 5 times faster ) .

Is it the only advantage of using Jacobian ? There are no more reasons than computaion time ? Like Idk , solution isnt much stronger/less sensitive to noise etc ???

Regards

Answer 1

[Torsten]

Stability of convergence is the main reason to supply an analytic Jacobian.

For computation time, it depends on the complexity of the analytical derivatives whether numerical or analytical Jacobian is more time-intensive. It may be the case that supplying an analytical Jabobian makes the computation slower than using the numerical alternative.