Unconstrained optimization is a basic tool in many disciplines, and there is no need to discuss its importance. This paper discusses how unconstrained optimization may be done using GNU Octave (Eaton, www.octave.org) using the package MINTOOLKIT (Creel, http://pareto.uab.es/mcreel/MINTOOLKIT). If you would just like to see some examples of how to use the algorithms, skip to section 4. Otherwise, here's some introductory information that explains how algorithms we selected for inclusion into MINTOOLKIT.