From Aesthetically Appealing Chinese Lattices to Routing or Circuitry Systems

Thomas Speller

MIT

Abstract

The purpose of this talk is to demonstrate the application of the shape grammar-cellular automata methodology (SGCA) to the ornamental artwork known as Chinese lattices (meanders), enabling computational generation of a catalog of designs in this style. This talk introduces the meander shape and describes the SGCA algorithmic stages of three shape grammar-cellular automata experiments on meanders. In the first experiment, a Chinese lattice specimen was selected as a model for creating a shape grammar, which was then transcribed into a cellular automaton that could generate the lattice style. The first experiment provided a fruitful subject for studying the variety of embedded shapes and symmetries in the meander system.

Following this first experiment, it was noted that by applying various constraints to a lattice, the meander could be interpreted as a routing or circuitry scenario. Thus, the second experiment built upon the first by incorporating a specification to the lattice and by working with more basic, generalizable meander primitives. This experiment was divided into two subparts, the first using the SGCA approach to generate enumeratively the creative space of systems based on the specification. The second subpart examined the specified meander’s computational space for understanding the characteristics of that space, affording a means to compare the use of certain evolutionary computation techniques against complete enumeration for managing the solution search. This comparative analysis led to guidelines for selecting efficient search techniques for different kinds of system architecture computational spaces and time constraints.

The third experiment extended the methods used in the second experiment to a real engineering system (underfloor heating systems), demonstrating an SGCA application for finding the most efficient system architecture solutions for a routing/circuit problem and providing an efficient (least action defined) select group of system architecture solutions from which stakeholders could choose.