TOMLAB /SOL

The TOMLAB /SOL toolbox efficiently integrates the well-known solvers developed by the Stanford
Systems Optimization Laboratory (SOL) with Matlab and Tomlab. The toolbox includes

• The sparse general nonlinear solver SNOPT 7.1-1,
   
• The sparse positive convex quadratic solver SQOPT 7.1-1,
   
• The dense general constrained nonlinear solver NPSOL 5.0,
   
• The dense constrained linear least squares and convex quadratic program solver LSSOL 1.05 and
   
• The dense constrained nonlinear least squares solver NLSSOL, based on NPSOL.
   
• Read more about the above mentioned SOL solvers, as well as the TOMLAB /MINOS solvers MINOS, QPOPT and LPOPT, at the SBSI-SOL Home page.
   
• TOMLAB /SOL includes all Matlab code and solvers that are part of TOMLAB /MINOS,
  see the description of TOMLAB /MINOS.
   
• For a general description of the Tomlab features, see the TOMLAB FEATURES.
   

  Main features

• The codes are implemented to efficiently handle Matlab sparse arrays.
   
• It is easy to use warm starts for the SOL solvers, and further speed up sequences of optimization solutions.
   
• NPSOL has been used intensively in control applications because of its speed and robustness.
   
• All SOL solvers may be used as sub problem solvers for other Tomlab solvers.
   
• The SPECS file input format, that is used by all SOL solvers when running non-Matlab, is also possible to use.
   
• NPSOL and SNOPT may estimate derivatives internally, faster than using any of the five Tomlab methods.