From the algorithm, we can see that it tries different damping factor until it gets a good one by the error. Is the damping factor related to the eigenvalues of the Hessian matrix?
$\begingroup$
$\endgroup$
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
The damping factor is expressed in terms of the objective function that one seeks to minimise, as described in A new approach for determining damping factors in Levenberg-Marquardt algorithm for solving an inverse heat conduction problem (2017).
answered Jul 5, 2019 at 10:10