From Meander Designs to a Routing Application Using a Shape Grammar to Cellular Automata Methodology

摘要

The usefulness of a methodology that integrates shape grammar for capturing design information with cellular automata for computational output of a design solution space is demonstrated in this paper. The application domain is the ornamental artwork known as Chinese lattices or meanders, a subject of earlier interest in shape grammar studies. In this study, a specification for a Chinese lattice is used for creating a shape grammar to capture the model's rules of self-organization, which are then transcribed into cellular automata to physically generate a catalog of designs that meet the requirements of this particular meander style. Then, the study compares the use of a probabilistic (evolutionary computation) technique against complete enumeration for managing the search for unique designs. In consideration of the finding of a very large number of rule solutions for a design specification which produced only a very small number of graphically unique architectures, the question is raised as to whether a more efficient search process other than brute force enumeration can be used. Finally, the meander study is extended to a real engineering system, demonstrating the applicability of the shape grammar to cellular automata (SG → CA) methodology for finding the most efficient system architecture solutions for a comparable routing/circuit problem. System architectures addressing an underground heating specification are automatically generated and evaluated, resulting in a group of design alternatives displaying the best piping layouts for the given requirements.