This would of course depend on the kind of optimization you are trying to do. If you have a global optimization problem with lots of local optima, perhaps ECJ's ES or PSO packages might do the trick. Otherwise, if your problem is really largely a basic gradient descent problem, ECJ might be too complex. At any rate, Liviu and I know ECJ well. Sean On Fri, 21 Jul 2006, Liviu Panait wrote: > Take a look at ECJ. You can find a link to it from the ECLab webpage. > > Liviu. > > On Fri, 21 Jul 2006, Michael Makowsky wrote: > > > I've posted this before, but I wanted to send out one more ping before I > > bought something. I am looking for a simple numerical optimization algorithm > > to use in one of my MASON models. I am currently trying to optimize a really > > easy quasi-concave function using a computationally expensive grid search. A > > simple numerical algorithm would really speed things up. > > > > > > > > Does anyone out there have some code or advice for me? > > > > > > > > mike > > > > > > > > > > > > Michael Makowsky > > > > Graduate Student > > > > Department of Economics, 3G4 > > > > George Mason University > > > > Fairfax, VA 22030 > > > > 703.608.1366 > > > > [log in to unmask] > > > > homepage <http://mason.gmu.edu/~mmakowsk> http://mason.gmu.edu/~mmakowsk > > > > > > > > >