OptLib is coming to sourceforge!

OptLib is a library of optimization routines that I've developed, mostly using stochastic methods, that I am in the process of putting on soureforge.

The library is currently in 3 parts, and I'm planning to write a unified interface to the three parts that will also allow for adding in more methods in a standard way. The 3 parts are genetic algorithms, simulated annealing, and Powell's method (Powell's method is deterministic, but I've included a Monte Carlo front end that uses a MC approach to guessing initial conditions to the Powell's solution--also I tend to use Powell's as a cleanup after each GA or SA run in practice.)

So far the GA routines are in the GIT repository, I'll add the SA and Powell's in the (hopefully) near future. The validation tests for each are all currently written into an educational model on optimization created for the Blue Water's project ( http://www.shodor.org/petascale/materials/modules/ -- listed under in development) which is undergoing review, and further GA validation tests are being worked into another paper in preparation for astrophysics applications. As soon as I have authoritative and reviewed references for the validation, I'll list it here as the preferred reference for anyone wanting to use this for academic purposes.