Home About Documentation Install Newsline Links Feedback Appeal Guestbook
Hello!
The page is obsolete, we have moved to www.openopt.org
OpenOpt citadel: Glushkov institute of cybernetics, NAS Ukraine
OpenOpt:
- New numerical optimization framework developed by optimization department of cybernetics institute (Ukrainian National Academy of Sciences; читайте також вступ українською мовою)
- Free Python-based cross-platform (Linux, Windows, MacOS etc) alternative to commercial frameworks like AMPL, GAMS, TomOpt's TOMLAB/TOMNET (see TOMLAB users and TOMLAB prices)
- Primary goals:
- providing connections to lots of solvers (some are C- or Fortran-written) with easy and unified OpenOpt syntax
- developing oofun syntax for non-linear problems group, that is capable to speedup some solvers (from or beyond SciPy) in several times ( take a look at 10X speedup example)
- trying to enhance algorithms for numerical optimization invented by ukrainian academicians Boris Pshenichniy and Naum Z. Shor
- providing connections to lots of solvers (some are C- or Fortran-written) with easy and unified OpenOpt syntax
- Framework license: BSD, connected solvers - any license (download URLs are provided)
- some native solvers (Python-written, NumPy required), constrained and unconstrained
- Convenient automatic gradients check for user-supplied objFunc and constraints
- Graphical output is available: examples 1, 2
