Generative Design With an Aesthetical Approach Using Multiobjective Optimisation

Abstract

Generative design provides an iterative process that allows for the optimisation of $3D$ models and is used in architecture to effectively explore a wide range of solutions and enable creativity in the design process. The evolutionary approach used in nature is imitated by the use of a test environment or artificial intelligence. A set of rules and constraints is determined for the optimisation to follow. Otherwise generative design is free from limitations and assumptions. Principles of beauty can be found everywhere, in nature as well as art and architecture, but also in mathematics and computational algorithms. It can be represented in ideas of symmetry, proportions and properties like the Golden ratio that can give a sense of order and consistency. In this thesis, the different ways generative design could be used to alter and optimise a design in accordance with external factors were explored. During a generative design process, both practical objectives, as well as aesthetical ones, were applied to a frame truss design that represented a stool. Strength and stability were examples of measurements for usability. The aim was a stool that was both practical to use and pleasing to the eye through the use of a multiobjective optimisation strategy. By proving that an optimisation that both considered practical and aesthetical objectives could be used on a simple frame truss design, it was presumed that this method could be used on more complex designs.