BOBYQA algorithm
This is a gateway function to use M.J.D. Powell's BOBYQA algorithm (Bound Optimization BY Quadratic Approximation) in MATLAB.
The C++ implementation of dlib (http://dlib.net) is compiled into a mex function to evaluate a MATLAB objective function.
The algorithm consits of four parts:
* The "bobyqa.m" MATLAB file contains the documentation, sets default values and calls the mex function (in .zip file).
* The "bobyqasub.m" MATLAB file is called by the mex function to evaluate the objective function which is MATLAB code (in .zip file).
* The "bobyqa_alg.cpp" file contains the C++ source code to create the mex function (in .zip file).
* The dlib source code containing the BOBYQA algorithm (download latest release at http://sourceforge.net/projects/dclib/files/latest/download).
The folder "/dlib-XX.YY" (XX.YY is your dlib version, e.g. dlib-19.16 for version 19.16) and the three files have to be copied into the same directory.
Run
bobyqa()
to display individual information on how to compile the mex function.
Further information as well as an example can be found inside the "bobyqa.m" file.
I have recieved several complaints that the first version of this was not working for some people (see comments). I cannot reproduce the described problems on my computer. If you encounter problems, please check your compiler settings (I used the MinGW-Add-On). If the problems persist, contact me and I will try to assist you in resolving your issues (might take a bit of time).
Comment: An analogous gateway function can probably be set up for the other C++ implementations if you find a good library.
인용 양식
Ulrich Römer (2024). BOBYQA algorithm (https://www.mathworks.com/matlabcentral/fileexchange/48689-bobyqa-algorithm), MATLAB Central File Exchange. 검색됨 .
MATLAB 릴리스 호환 정보
플랫폼 호환성
Windows macOS Linux카테고리
태그
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!