Levenberg–Marquardt Algorithm (LMA) is considered most
powerful iterative solution strategy for the optimization applications.
This algorithm is a blend of Gradient Descent Algorithm
(GDA) and Gauss–Newton algorithm (GNA). LMA proceeds with
the GDA, due to its low sensitivity to initial values. Once GDA calculates
values near to the solution, GNA automatically takes over
the process and accelerate its convergence.