Open source optimization software
If you are able to afford commercial license, CPLEX, AIMMS, ILOG, FICO xpress, LINGO, AMPL (with free student version, 300 variable, 300 constraints), LOQO. Open source packages also can solve some of optimization problems (NOT all).
(1) Lpsolve, mainly used to solve mixed integer problems, manual. It can solve problem as large as 30, 000 variables and 50,000 constraints, source. It also has JAVA version. LPsolve can be used on Windows, Mac, and Linux.
(2) GLPK, used to solve mixed integer and linear programming problems. It has nice documentation. It can also be run on different OS platforms. GLPK can be used to solve large scale problem, although there is not a rigious test for the scale yet.
(5) LOPTI, similar to OOL, it provides four solvers NEWUOA, CONDOR, Nelder-Mead (simplex) and Hook-Jeevs to solve the constraint optimization problems.