SHAPE GRAMMAR DESIGN ASSISTANT IMPLEMENTATION ON SPACE CONFIGURATION DESIGN

• Main Author: F P Sondakh, Alva
• Format: Theses
• Language: Indonesia
• Subjects: Arsitektur
• Online Access: https://digilib.itb.ac.id/gdl/view/36307
• Institution: Institut Teknologi Bandung

Space configuration designing is one of the core creative activity in a design process. In this phase, an architect creates a design as a respond to the need of the user within the constraints of design criteria. This creative skill is required from every architect, but in a high complexity project space configuration design process becomes very complicated that architects utilize computational methods in aiding them. One of the computational methods used is the Shape Grammar method. This method allows designer to heuristically solves the design problem by choosing the shapes and their rules to be used in creating the design. Many research has been conducted to the theories, concepts, and methods of Shape Grammar. One of which is the development of computer programs that take the task of Shape Grammar Design Assistant using Textual Programming Language (TPL) or Visual Programming Language (VPL). VPL has become a better choice than TPL because of its relative implementation in constructing a computer program aiding architects in their creative activity. However, in constructing a Shape Grammar Design Assistant an architect must still give a great attention in the process of developing the program’s algorithm. The development of the algorithm is not an easy task for architects as they are not computer programmer by training. The ideal condition is the tasks in a Shape Grammar Design Assistant program are automated so architects can focus on the creative process of constructing the Shape Grammar rather than constructing the program itself. This research is then conducted to explore the possibility of a parametric modular visual algorithm in supporting an architect to construct a Shape Grammar as a basis in spatial configuration creation. The exploration will be done through 4 phases in a type 3 Design Research Methodology by Blessing and Chakrabarti. In general the purposes of this research are to construct an algorithm that function as a Shape Grammar Design Assistant, then to explore its implementation in creating variations of residential space configuration design, and then lastly to evaluate the performance of that algorithm in terms of its effectiveness and its efficiency. The main problems of this research are what kind of shape grammar the algorithm will assist to design, how to construct the algorithm, how to use it, and then what kind of impact that the algorithm will bring to a design process. In answering those problems, this research will be limited to a case study of residential space configuration design with study object of a few residential works by Adolf Loos that represent the characteristics of Raumplan Shape Grammar. There are some notable output and findings from this experimental exploration that answer the research problem formulated in the beginning. The first output is a Shape Grammar Design Assistant algorithm inside a cluster of components in Grasshopper VPL Editor. The cluster called ‘RUPA’ contains an algorithm of 2 or 3 dimensional shape creation and transformation (i.e. translation, rotation, reflection, dilatation, and strain. Using that cluster a user can create and transform 2 or 3 dimensional shapes by changing the value of its parameters. However, a potential major drawback will emerge in creating a complex space configuration design because of the large number of clusters that would be put into a single program or definition. A further research can be done to optimise the algorithm inside the cluster so that the execution of the program can be faster and efficient. The next output of this research is a method of contructing a Shape Grammar which proposes a concept of Reference Geometry and Result Geometry. In this method, a Reference Geometry is used as a platform or a host geometry for the creation and tranformation of Result Geometry. That Result Geometry then can be used as a Reference Geometry in the next iteration of the process to completely construct the Shape Grammar. This method has been successfully implemented in creating space configuration design based on the Shape Grammar of some residential works by Adolf Loos. However, the method is still in early stages of development, and needs to be tested to other design models to evaluate its effectiveness and repeatibility. The consequent output of that method is a Shape Grammar of Adolf Loos residensial works constructed in clusters of ‘RUPA’. The ‘RUPA’ construction was done according to one of two type of clusters generation, i.e. the Dependent type. It was discovered that Dependent type have made the process of Grammar construction easier because other than arranging the clusters the user only have to change the value of parameters in a or some Key Shapes to create the grammar. Even so, it was also discovered that the process of construction using the Dependent type requires the user to have adequate knowledege of arithmetics and programming logic. Other than those primary findings related to the research problem, this experimental exploration is successful in creating space configuration design and its derivation based on the Shape Grammar of residential works by Adolf Loos. The creation process began with an analysis phase with its methods experimentally developed specifically for this research. The reference Loos residential works were analysed and reverse engineered using the concept of Reference Geometry and Result Geometry. The results were satisfactory, and were used in creating design derivations of each analysis objects. The analysis phase took a significantly large amount of time relative to the whole time took to conduct the research. It was because the analysis phase must be done manually to each analysis object. The derivation generation phase was relatively simple because the user only need to change the value of parameters of the Key Shapes to find the suitable design derivation. However, the changes given to the parameters must be within a rational range of values so the resulting shapes is not too big or too small. Further potential researches can be conducted to develop the outputs of this research, among which are researches that look into how to simplify or even automating the analysis phase of shape grammar of existing design using Artificial Intelligence or Machine Learning technology. Another potential research is a research that looks into further interpretation of the ‘RUPA’ generated Shape Grammar for better understanding a certain body of architectural works.