OPTIMIZATION IN PARAMETRIC DESIGN SPACE

Optimisation of complex design configurations in terms of performance is rarely achieved due to the computational effort required to explore multi-dimensional design space. Genetic algorithms (GAs) are currently being employed by many research groups for that purpose but tend to be configured as software systems which are independent of the CAD modelling system. This paper describes a CAD system where design for product performance, design for manufacturing and design for cost are accommodated within one integrated adaptive environment. One aspect of this system is covered in depth in this paper, the evolutionary optimisation of performance and cost using hybrid coevolving GAs.