ABSTRACT
ParaGen is a method for exploring solution options in the early phases of design. The method is intended to aid the designer by exposing solutions which perform well against an array of performance criteria. ParaGen makes use of a variety of commercial simulation software to determine the desired performance parameters in relevant areas such as structures, lighting, acoustics or energy. Based on the values of these parameters a multi-objective search is carried out using a Non-Destructive Dynamic Population Genetic Algorithm (NDDP GA). In this approach all solutions, including relevant performance parameters as well as geometric data are retained in a relational database both as quantitative values and as qualitative images. SQL queries are used both by the NDDP GA to guide the search and by the designer directly to explore the solutions. A number of tools are also included to aid the designer in finding well performing solutions. These include graphing functions to find Pareto sets of conflicting performance criteria and parallel coordinate graphs to examine the interrelation of model parameters and performance values.
