Back to results list
Please use this identifier to cite or link to this item:
|Title:||The development of parametric shape grammars integrated with an interactive evolutionary system for supporting product design exploration||Authors:||Lee, Ho-cheong||Keywords:||New products
Hong Kong Polytechnic University -- Dissertations
|Issue Date:||2007||Publisher:||The Hong Kong Polytechnic University||Abstract:||This thesis has developed an interactive system that uses parametric 2D and 3D shape grammars incorporating an evolutionary algorithm for exploring product forms at the early stage of design process. The evolutionary algorithm allows designers to explore product forms by modifying the control parameters of shapes and changing the application of shape grammar rules based on the interaction between designers and the system. In recent years, shape grammars have been used in Computer-Aided Design (CAD) systems for generating stylistically consistent and novel designs. In this approach to design concept generation, human subjective selections and evaluations are involved in controlling the parametric modification of the shapes and shape grammar rules. System based on this approach is referred to as Interactive Grammar Based Design System (IGBDS). An IGBDS is capable of reproducing large numbers of stylistically consistent designs as well as exploring novel designs. The design approach based on shape grammars, however, relies on the identification of quality shape grammar rules with sufficient accuracy but in the meantime maintaining a high degree of generality, in order for the system to generate forms with enough variations. The formulation of shape grammar rules useful for different specific design requirements in a computational system is difficult and time consuming in the domain of product form design, due to the complexity in the product surface modelling. This is so because the development of shape grammar rules requires tremendous amount of work in capturing the knowledge from the existing designs. However, most of the existing designs have different form features. It is difficult to find the consensus in generalizing the existing designs into shape grammar rules. In run time, the selection and evaluation processes of the shape parameters and the shape grammar rules are time consuming due to limited speed in rating and interaction between designers and the computer. To address these problems and to enhance the generative capability of shape grammars in supporting product design, this research focuses on two issues: 1) The development of systematic approach to the formulation of shape grammars combining 2D and 3D forms and 2) Extending the generative capability of the product design support systems which use shape grammars. The aim of these two focuses is to enhance the power of shape grammar based design approach in terms of dealing with complexity of real design applications and to explore the potential of integrating evolutionary methods with shape grammars for increasing diversity in product design. The motivation for such an integration is based on the hypothesis that shape grammar rules modified by the genetic code scripts of an evolutionary method (such as genetic algorithm) define a new combination of shape features for alternative designs. In this way, traditional shape grammars are extended to an interactive context in which generative and evolutionary computing methods are utilized with better potential of supporting product component design and configuration. Different scenarios in which designers interact with the implemented system in real design applications are studied and evaluated. The first issue involves an analytical approach to understanding the relations between product design and form complexity, by identifying and analysing the information associated with complex form creation. The research on this issue included the development of methods for the classification of product components, the definition of design spaces for the component configuration, the specification of design constraints and the spatial relationships among components.
The second issue involves the development of a computational system for enhancing the generative capability of IGBDS in the product form exploration, by integrating an evolutionary algorithm with an IGBDS. This integration allows shape grammar rules to be modified at run time. This relied on a careful planning in retrieving the significant elements from the complex information network of shape grammars and encoding all the necessary information into genetic representation. A new representation scheme utilising both the power of genetic programming and shape grammar rule representations is developed, together with the control strategies for manipulating this new representation during an evolutionary design process. With such a method and system, the designers come to interact with the enhanced IGBDS to evaluate the designs and corresponding rules by determining the control parameters and control strategies in the evolutionary cycle. The system provides the visualisation of virtual 3D objects for the designers for their evaluation and selection, and allows them to explore alternative designs with shape grammar rules containing parametric variation potentials, through the integration of Genetic Algorithm (GA), Genetic Programming (GP) and parametric 2D and 3D shape grammars with labels. The methodology for research carried out in this thesis follows the analytical study of shape grammars and product forms with the development of new computational representations for the integration of evolutionary methods with shape grammars. The implemented system and knowledge base of the shape grammars are tested with realistic design examples. The research involves three major steps: 1) The development of the methods of deriving shape grammar rules in a product design domain, 2) The formulation of genetic representation and control strategies, and 3) The exploration of design and evaluation of the integrated IGBDS with real design examples. With the system developed in this research the designers can explore design with more flexibility in varying the weighting factors, determining the control parameters for the shape grammar rules, selecting appropriate control strategies for specific design characteristics, controlling order sequences of shape grammar rule application, and exploring form and configuration designs. For the evaluation of the proposed method, representation and implementation of an enhanced IGBDS, a shape grammar rule base for digital camera design is developed and two prototype systems are implemented using this rule base in order to study the feasibility of proposed system with its supporting components, including the integration with an external 3D solid modelling kernel. The development and experiments of two prototype evolutionary IGBDS illustrated the flexibility in using parametric 2D and 3D shape grammars with labels to explore a wide range of engineering and industrial design problems involving complex and intuitive conception of forms and products. Both implemented systems are interactive, with the first one using a normal evolutionary algorithm, and the second one using genetic programming with parametric shape grammars. The experiments showed that the second prototype is more flexible, while more complex to implement, for the exploration of forms in terms of variety with the parameters within shape grammar rules being changed by the evolutionary algorithms. Based on the analysis and the evaluation of the results achieved in the context of real product design, the thesis concludes on the applicability of shape grammars to real product design and proposes several strategies with which this research can be further advanced for complex form design and visualisation in product design.
|Description:||xvi, 201 leaves : ill. ; 30 cm.
PolyU Library Call No.: [THS] LG51 .H577P SD 2007 LeeH
|URI:||http://hdl.handle.net/10397/2256||Rights:||All rights reserved.|
|Appears in Collections:||Thesis|
Show full item record
Files in This Item:
|E-thesis_Link.htm||For PolyU Users||161 B||HTML||View/Open|
|b21458947.pdf||For All Users (Non-printable)||4.68 MB||Adobe PDF||View/Open|
Citations as of Oct 15, 2018
Citations as of Oct 15, 2018
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.