OPTIMIZATION WORKSHOP
A group of the MAA students had the chance to participate in the workshop tutored by Sam Conrad Joyce (Buro Happold) aimed to employ metrics to quantitatively evaluate the performance of a design and to use Genetic Algorithms to optimize a design, in general. The session was divided in two parts: a theoretic introduction to the optimization processes of the design and the concept of Genetic Algorithm and an applicative part where the participants optimized their own design with the support of the tutor.
Parametric design principles are being employed more and more. Whilst offering flexibility and possibility, these methods ultimately require decisions to be made to confirm the final form. A practical paradigm to pursue is performance driven design, where parameters are set based upon required performance, whether this is geometrical (proportion), environmental, structural or economic performance. Naturally this is not trivial, as no single algorithmic method can improve/optimize every design. Of special interest are iterative Newtonian type solvers, and Genetic Algorithms. The tutor has experience of applying these in practice (Buro Happold), in projects such as the design of the Louvre in Abu Dhabi, and more generally for individual structural components through to larger scale master planning schemes.
The workshop aim was to cover the following topics;
• How to employ metrics to quantitatively evaluate the performance of a design
• Methods of mapping performance to multi-dimensional parameter spaces as well as understanding sensitivity analysis
• Using algorithms to optimize a design
