This project began more than a decade ago as part of the dissertation of Bill Hart, focused on the development of powerful global optimization techniques combining evolutionary algorithm sampling methods with gradient-directed local search. These hybrid techniques proved useful as a component of the AutoDock system developed by Art Olson and colleagues at Scripps Research Institute. In this application, the goal is to search for minimal conformation energies of small molecules into receptors of known structure. AutoDock has since proven useful to many as a component of rational drug design efforts targeted especially towards HIV protease inhibitors. The second phase of this project integrated the co-evolutionary design techniques developed as part of Chris Rosin's dissertation in 1997. These experiments considered not only the efficacy of drugs with respect to "wildtype" virus, but also their robustness across a wide range of the mutants that can arise in this quickly evolving system.