Article ID: | iaor20133781 |
Volume: | 48 |
Issue: | 1 |
Start Page Number: | 187 |
End Page Number: | 199 |
Publication Date: | Jul 2013 |
Journal: | Structural and Multidisciplinary Optimization |
Authors: | Keane Andy, Nasuf Alkin, Bhaskar Atul |
Keywords: | heuristics: genetic algorithms |
We propose an automated shape generative framework, which provides an alternative way of exploring the design space in a structural mechanics context. The framework presented uses ‘blind’ evolutionary intelligence to synthesise shape grammar sentences i.e. Grammatical Evolution (GE), where rules are selected by a Genetic Algorithm (GA). This is a novel approach to automate the Shape Grammar (SG) formalism. We then present an application of a grammar based shape generative framework to solve a 2D design optimisation problem. This involves synthesis of parametric 2D curves where the shape grammar primitives are introduced as arcs represented by rotation and a radius. The efficacy of the proposed shape generative framework is then compared with that of Non‐Uniform Rational B‐Splines (NURBS) parametrisation for structural optimisation.