Automated Induction of General Grammars for Design

Grammars are useful for representing systems of design patterns, however formulating good grammars is not straightforward in many contexts due to challenges of representation and scope. This challenge has been identified as one of the 3 goals for computerized use of shape grammars: grammar inference. This work introduces a highly flexible mechanism for inducing viable general grammars from a computational representation of a designed context. This mechanism is evaluated with several common types of devised media of increasing complexity based on dimensionality: 1D (e.g., text), 2D (e.g., PCB layout, building plans), many dimensional (which in abstract can generally be used to represent product, system, platform or service designs), and, against a set of grammar properties necessary for a grammar acquisition method to be useful: accuracy, variability, repeatability and conciseness. This work shows complete enumeration over possible grammars in the 1D case and a continuum of approaches for higher dimension data sets that are demonstrative of grammars in design .