Splitting Metrics Diagonal Bundle Method
"Results! Why, man, I have gotten a lot of results. I know several thousand things that won't work."
- Thomas A. Edison
SMDB
SMDB is a Fortran95 implementation of the splitting metrics diagonal bundle method for general, possible nonconvex, nonsmooth (nondifferentiable) minimization. The software is free for academic teaching and research purposes but I ask you to refer to the reference given below if you use it.
Code
| tsmdb.f95 | - main program for SMDB. |
|---|---|
| initialization.f95 | - user-specified initial values for parameters and starting point x. |
| objfun.f95 | - user-specified computation of the objective function value and the subgradient. |
| parameters.f95 | - modules of parameters. |
| smdb.f95 | - SMDB. |
| subpro.f | - subroutines for SMDB method. |
| Makefile | - makefile. |
| testps.f95 | - module of large-scale nonsmooth test problems (includes also partially separable test functions). |
| smdb.tar.gz | - all the above in compressed form. |
To use the software modify tsmdb.f95, initialization.f95 and objfun.f95 as needed.
References for SMDB
- N. Karmitsa, M. Gaudioso, K. Joki: "Diagonal Bundle Method with convex and concave updates for Large-Scale Nonconvex and Nonsmooth Optimization". Submitted, 2017.
- N. Karmitsa, M. Gaudioso, K. Joki: "Splitting Metrics Diagonal Bundle Method for Large-Scale Nonconvex and Nonsmooth Optimization". TUCS Technical Report, No. 1178, Turku Centre for Computer Science, Turku, 2017.
Acknowledgements
- The work was financially supported by the Academy of Finland (Project No. 289500).
