First, if you desparately need to solve a nonsmooth optimization problem, you have come to the right site. Otherwise, you still may use SolvOpt, however, it might happen you have missed a better choice.
As it was stated, SolvOpt provides an optimization tool for almost every problem, however, the weak point of the code is seen in its general applicability. The latter means that it cannot compete in efficiency with the optimization codes designed specificly for a particular class of problems. It is senseless to use SolvOpt for solving a linear programming (LP) or convex quadratic programming (QP) problem. There are many other optimization packages available via Internet, which are much more efficient in the mentioned application areas. If you still doubt, you better visit the sites: http://plato.la.asu.edu/guide.html (The Guide to Optimization Software by Hans D. Mittelmann) and http://solon.cma.univie.ac.at/~neum/glopt.html (Global (and Local) Optimization by Arnold Neumaier) providing the most complete list and comprehensive introductions to the existing packages. You may also visit Netlib for even more information. There are many more sites worth to visit, but we do not see our goal in providing the most complete list of links.
Since the time the first version of SolvOpt has become available freely, the authors have been receiving many comments and requests. This is why we have decided to provide the FAQ section.
Online reference, Performance and tests.
Alexei Kuntsevich ||
& Franz Kappel |
June 15, 1997