The project brief calls for the design of an office building located in Tokyo. One of the agendas of the studio was to generate architecture algorithmically using the Processing platform. To that end, my intention was to define local intelligence at a granular scale to control tectonics on a larger scale. The technique was utilized as a means of form generation. In other words, I was interested in controlling the brush stroke by defining the brush.
A cube was given a set of simple behaviors to follow. If two cubes overlapped, they would move away from each other. If a cube detected that there was free space in any direction, it would grow in that direction. Initially, hundreds of cubes were instantiated all in the same position. Their behavioral intelligence allowed the system to quickly develop a high level of complexity and spatial variety. Next, the cubes were instantiated along a random noise path, restricted by the boundaries of the site. This allowed the system to fill in the spatial envelope of the program. To give some variation to the homogeneous system, an attraction behavior was introduced. While the cubes are being generated in one of three colors, corresponding to public space, office space, and leased space, a given percentage of the cubes are designated as “attractors”. These cubes pull other cubes of identical color towards them if within a given radius. This creates a system that consists of large regions of cubes delineated by color. Once all the cubes have been attracted, the grow behavior is enabled. Lastly, a gravity system is introduced that pulls the attractor cubes (along with the rest of the cubes) down towards the ground.
Although the behavioral concepts described are quite simple, this technique permitted the creation of numerous iterations of the project that were both formally complex and intelligent in their organization. Spaces generated ranged from intimate personal work areas, to large volumes suitable for public events, all articulated by cubes of various scales.