I have seen the implementation of L-BFGS-B by authors in Fortran and ports in several languages. I am trying to implement the algorithm on my own.
I am having difficulty grasping a few steps. Is there a worked out example using L-BFGS or L-BFGS-B ? Something similar to (attached link) explaining the output of each step in an iteration for a simple problem.
http://www.cas.mcmaster.ca/~cs4te3/tutorials/BFGS.pdf
Look an implementation in MAXIMA. "MAXIMA.sourceforge.net". The code procedure is open and free. Due to the clear MAXIMA language the code is easy to understand.
You might be interested in
https://github.com/scipy/scipy/blob/master/scipy/optimize/optimize.py
Thanks all for the links. I was ableto implement the LBGFS, by reading the paper from Nocedal : http://www.ii.uib.no/~lennart/drgrad/Nocedal1980.pdf Still trying to implement LBFGS-B
@Vladimir Kobolev : I found some LBGFS related material here for MAXIMA http://cs.swan.ac.uk/~csoliver/ok-sat-library/internet_html/doc/doc/Maxima/5.29.0/html/maxima_61.html but nothing related to LBGFS-B, which is for the bounded case.
@Nils Wagner : Optimize.py has LBGFSB, am still trying to figure it out. Any help text or explanation besides the comments are appreciated.
@Joachim Pimiskern : DLIB code is very readable and helpful, but I think it only does LBGFS and not LBGFS-B, i.e the bounded case.
Thank you for your reply.
@Joachim Pimiskern : I think you are suggesting the penalty method, and we have been trying that too, but to my experience, it has been much slower. What are your thoughts about the comparison of speed between BFGS with penalty method vs LBFGSB ?
@Shawkat Hamdan : Do you mean Particle Swarm Optimization by PSO ? In that case, are you suggesting to replace LBFGS-B with PSO or augment one with another ? Can you provide more information about the same ?
yes i mean Particle Swarm Optimization by PSO ,yes i suggest to replace LBFGS-B With PSO .i solve this function using PSO: Minimize f(x) = exp(x1-1)+ exp(-x2+1) + (x1-x2)^2 and i get :x1=0.7962,x2=1.204 and minimum of f(x) =1.7974 have you another objective function to solve with constraints or without constraints ?
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Algorithms