Formalization and meaning of the architectural project
Applicants: Gilles Falquet, Pierre Pellegrino, Daniel Coray
Researchers: Emmanuelle Pellegrino Jeanneret, Krishnendra Shekawat, Mathieu Vonlanthen
This follow-up project has two main aims: 1) to integrate the tools and techniques developed in the initial project into a common extensible and open architectural project environment; 2) to develop new theoretical results to deal with important issues such as flexible constraint resolution, open and closed world reasoning, combining spatial reasoning and inference rules, or finding new relations between architectural co-variants.
Our work on algorithms for shape generation and constraint resolutions, logical rea- soning, ontological definitions, etc. has clearly shown that a architectural project system must encompass a large body of creation and resolution tools. Thus it appears necessary and highly interesting to extend our information management system for the architec- tural project to an open and extensible environment for architectural project. The first use of this environment will be to integrate the algorithms and tools developed in the initial project.
The architectural project environment will be built around a common knowledge base that represents general architectural knowledge, knowledge relative to a particular style, and factual knowledge about the current project. Some part of the knowledge will also be stored as inference rules and algorithms. A computational layer will provide reasoning services such as subsumption and instance checking, abduction, rule evaluation, etc. , together with project-oriented services : shape generation, constraint resolution, semantic classification. The user interface will be comprised of a logical editor (to work on the ontological part of the knowledge base) and an interactive drawing tool (AllWright). This tool will enable a designer to create and update the components (syntax) of a project and to invoke computational services to reason about the current project or propose solutions to specific problems.