Levenberg Marquardt method using Golden Section algorithm

19 views (last 30 days)
Andrew Green on 22 Dec 2021
Answered: arushi on 13 Feb 2024
I could not run the Levenberg Marquardt algorithm with the golden section method. I'm new to Matlab and I've read a lot of articles but I couldn't find any similar examples. Is this not possible?

arushi on 13 Feb 2024
Hi Andrew,
The Levenberg-Marquardt algorithm and the Golden Section search method are two different optimization algorithms that are generally used for different types of problems.
The Levenberg-Marquardt algorithm is typically used for non-linear least squares problems, which are common in curve fitting and data fitting scenarios. The algorithm combines the gradient descent method and the Gauss-Newton method to minimize a sum of squared function values.
On the other hand, the Golden Section search method is a line search method used to find a local minimum of a unimodal function in one dimension without the use of derivatives.
Combining these two methods is not a standard approach because they are designed for different optimization tasks. The Levenberg-Marquardt algorithm already includes a mechanism for adjusting the step size, which is somewhat similar to the goal of the Golden Section method in finding an optimal step size in one-dimensional problems.
In MATLAB, if you want to use the Levenberg-Marquardt algorithm, you can use the “lsqcurvefit” or “lsqnonlin” functions from the Optimization Toolbox, which have the option to specify the Levenberg-Marquardt algorithm as the optimization method.
Please refer to the MathWorks documentation for “lsqcurvefit” function –
Please refer to the MathWorks documentation for “lsqnonlin” function –
Hope this helps.

Categories

Find more on Linear and Nonlinear Regression in Help Center and File Exchange

R2021b

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!