image thumbnail

global_optim_fittin​g_matlab

version 1.0.0 (91.9 KB) by Thomas Guillod
MATLAB Toolbox for Global Fitting/Optimization

20 Downloads

Updated 2 Jun 2022

From GitHub

View License on GitHub

MATLAB Toolbox for Global Fitting/Optimization

license - BSD language - MATLAB category - science status - maintained

This MATLAB toolbox can be used for the following problems:

  • finding global minimum of a function
  • fitting a function to a dataset

This toolbox is specially adapted to the following problems:

  • non-smooth error function
  • non-convex error function
  • computationally heavy error function
  • error function with local minima
  • error function with many input variables

This toolbox provides a common interface for different solvers:

  • gradient: fminunc / fmincon
  • simplex: fminsearch
  • surrogate: surrogateopt
  • evolutionary: particleswarm / ga
  • the aforementioned solvers can be combined

Customized error function:

  • custom weights for the dataset points
  • choice of the error metric (norm, average, percentile, etc.)
  • recover from undefined values
  • vectorized evaluation of the error function
  • parallel evaluation of the error function
  • caching of the error function

Advanced variable handling:

  • abstraction layer for the variables
  • initial values
  • scalar or vector variables
  • variable transformation (linear, quadratic, logarithmic, etc.)
  • variable normalization
  • constraints (lower and upper bounds)
  • sine transformation for handling constraints

Advanced monitoring capabilities:

  • compute various error metrics
  • compute solver figures of merit
  • plot/display the solver progress
  • plot/display the final results

Limitations

  • All the provided features have a computational cost.
  • Therefore, this library is mostly adapted to time-consuming error functions.
  • For simple error functions, the overhead is non-negligible.

Examples

Compatibility

  • Tested with MATLAB R2021a.
  • The gads_toolbox is required (for the MATLAB solvers).
  • The optimization_toolbox is required (for the MATLAB solvers).
  • The distrib_computing_toolbox is required (for parfor loops)
  • Compatibility with GNU Octave not tested but probably problematic.

Author

Thomas Guillod - GitHub Profile

License

This project is licensed under the BSD License, see LICENSE.md.

Cite As

Thomas Guillod (2022). global_optim_fitting_matlab (https://github.com/otvam/global_optim_fitting_matlab), GitHub. Retrieved .

MATLAB Release Compatibility
Created with R2022a
Compatible with any release
Platform Compatibility
Windows macOS Linux

Community Treasure Hunt

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

Start Hunting!
To view or report issues in this GitHub add-on, visit the GitHub Repository.
To view or report issues in this GitHub add-on, visit the GitHub Repository.