Categorisation of Designs According to Preference Values for Shape Rules

Abstract

Shape grammars have been used to explore design spaces through design generation according to sets of shape rules with a recursive process. Although design space exploration is a persistent issue in computational design research, there have been few studies regarding the provision of more preferable and refined outcomes to designers. This paper presents an approach for the categorisation of design outcomes from shape grammar systems to support individual preferences via two customised viewpoints: (i) absolute preference values of shape rules and (ii) relative preference values of shape rules with shape rule classification levels with illustrative examples.

Shape Exploration in Design: Formalising and Supporting a Transformational Process

“…Analysis of experimental data produced a number of general/detailed shape rules that formalise the shape transformations used by designers when exploring design concepts.Although the rules express shape transformations in an abstract way without representing the exact transformation of a shape, the analysis suggests that the interactions of designers with their sketches can be represented by a finite number of shape rules.The rules formalise the reinterpretation and transformations of shapes, e.g. through deformation or restructuring.The shape rules defined in this research are used to inform development of a computer-vision based shape grammar system, developed to implement our research.The system offers a fluid interaction with digital representations of design that reflects the modes of interaction observed in designers as they explore design concepts via sketching.The system also provides the potential to generate designs within design families via application of shape rules.The analysis reveals a possibility of a customisable selection of generated design outcomes which might not only improve shape grammar system performance but could also provide more meaningful outcomes to designers based on personal design intentions.The support of personal design intention with customised viewpoints that use hierarchical classification of shape rules with preference value for each shape rule is described in [29]. Future work is concerned with exploring (i) how the defined shape rules can be further detailed in a hierarchical manner, (ii) how the customised viewpoints that have been formalised can be further developed, and also (iii) how these results can be further integrated with the computational tool for conceptual design that has been developed for our research.…”

“…If a designer considers that the shape rule R a is most important to cluster an object, then the shape S 1 and S 2 could be classified in the same cluster. In all other cases, they would be classified in a different cluster.This can frequently happen when a designer and user are different.The customisable selection via different criteria might not only improve shape grammar system performance but could also provide more meaningful outcomes to designers [29]. It is suggested that this can be done by parameterisation of shape rules adapted from the method for numerical representation of vagueness [28], which parameterises vague geometric ᭤ Table 1.…”

A Formal Approach to the Study of the Evolution and Commonality of Patterns

“…Therefore, the hierarchical deconstruction of an artefact into meaningful components is necessary to defi ne counterparts which can be compared with other artefacts in different styles. Despite shape grammar's potential to generate infi nite designs using a predefi ned set of shape rules, shape grammars require a hierarchy of shape rules, to generate designs more systematically (Lim et al, 2008b). It is because the systematic comparison of the counterparts allows the use of shape grammars in the interpretation of artefacts with a strong awareness of metacontents, like meaning embedded in the counterparts.…”

“…It also allows a formal investigation comparing the geometric elements of the counterpart. However, various hierarchical deconstruction methods have been successfully employed to synthesise and analyse the metacontents of an artefact in the fi elds of molecular biology (Averof and Patel, 1997;Carroll et al, 1995), anatomy (Chen et al, 1994), pathology (Victoria et al, 1997), computer graphics (Seo et al, 2010), and design optimisation (Choudhary et al, 2005;Lim et al, 2008b;Scheepers et al, 1997).…”