A self organizing map (SOM) is a model of a artificial neural network developed by the Finnish professor Teuvo Kohonen. A SOM is used to rearrange the plots of the inner city of Zurich so that similar plots are close to each other. Continue reading
An algorithm generates an endless alphabet of floorplan-variations. The resulting plans are clustered by similarity. Topological and geometrical properties are iterated, from 1-room apartments to 4-room apartments.The resulting plans are sorted by a self organising map (SOM) to cluster similar compositions together. Continue reading
An information retrieval system for parcels is introduced that not only allows searching for specific attributes, but also includes properties of shape and context of the building plots.
Increasingly, digital architectural data will become available through information technology. Yet until now, there were no satisfying methods to query this data for architectural purposes. An information retrieval system for parcels is introduced that not only allows searching for specific attributes, but also includes properties of shape and context of the building plots. An automatically generated index stores the relevant spatial properties as normalized bitmap images on several layers. When a query is started, only this index has to be queried and not the complete database. The search process can be controlled through a graphical interface that incorporates the user’s sketches. The retrieved parcels are presented as a sorted list of vector drawings including their contained buildings. With the simplified access to these case-studies, quality and efficiency of the architectural design process could be increased. Continue reading
A library of floor-plans with over 1000 appartments of existing houses. The database allows complex search-queries by topological and geometrical properties, concerning building-elements, rooms, and organisation. Continue reading
