Many design tasks involve selection from a set of configurations followed by parametric optimization of the chosen configuration. The models used for these tasks tend to be large, non-linear and involve both discrete and continuous variables. It is rarely possible to use any single formal algorithm to solve these problems and as a result there are few tools to help designers solve such problems. We believe computer environments that allow flexible access to a varied set of tools will help designers rapidly generate high quality solutions. We demonstrate our arguments on a design problem taken from a commercial auto manufacturer, propose a framework for dealing with the general class of problems and describe an implementation of a novel design system that integrates math programming, knowledge-based and graph theoretic tools.