optimization with Levenberg-Marquardt algorithm

9 次查看（过去 30 天）
ghazal mohamed
ghazal mohamed 2018-1-23
评论： ghazal mohamed 2018-1-24
Hi everyone! I have a function f(x) (ex: f(x)=3+sin(2x)-exp(-x)) and it's experimental value Fexp (ex:0.02344). So how I can find x that minimize the difference f(x)-Fexp using Levenberg-Marquardt algorithm in Toolbox optimization? Many thanks
  2 个评论
Matt J
Matt J 2018-1-23
Levenberg-Marquardt would be overkill. Since it's a 1D problem, you could just use fzero().
ghazal mohamed
ghazal mohamed 2018-1-24
hi Matt,thank you for your comment, the function given is just an example, the expression I want to minimize is more complicated than that. On the other hand I am not looking for zero of function, simply because f(x)-Fexp (in my case) dosen't have zero so I am looking for x that gives the smallest value of f(x)-Fexp using Levenberg-Marquardt algorithm.

请先登录，再进行评论。

请先登录，再回答此问题。

回答（1 个）

Alan Weiss
Alan Weiss 2018-1-24
编辑：Alan Weiss 2018-1-24
I am not sure that I understand you, but perhaps you are asking about Curve Fitting Via Optimization.
You are free to use the Levenberg-Marquardt algorithm in lsqcurvefit or lsqnonlin when fitting a nonlinear function to data. See the Algorithm option.
Alan Weiss
MATLAB mathematical toolbox documentation
  1 个评论
ghazal mohamed
ghazal mohamed 2018-1-24
Hi Alan,thank you for your reply, I am not looking for data fitting. I have function f(x) and an experimental value Fexp, if I consider g(x)=f(x)-Fexp, x is a variable in [0 100], g(x) takes differet values depending on x, among these values (of g) there is a minimum :gmin, gmin is obtained when x=Xm so I am looking for this value Xm. (I think that I can't use a nonlinear solver because g(x) has not zero but there is a minimum value..).

请先登录，再进行评论。

类别

在 Help Center 和 File Exchange 中查找有关 Choose a Solver 的更多信息

标签

产品

Community Treasure Hunt

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

Start Hunting!

Translated by