This video demonstrates some of the uses of generative algorithms as applied in urban context. Most of them operate on agent based systems (sometimes multi-layered), and exhibit low-level self organization qualities, like clustering.
It roughly outlines most of computational design work I've done within last few years.
This is a prototype of a table simulating pedestrian flow from three tube stations to three separate retail/office entrances. Agents are given a task of reaching randomly chosen entrance and returning to randomly picked tube station. Thus, basic crowd mechanics is achieved... Experiment takes place in London (Cannon Street area).
Three buildings of arbitrary shapes can be physically placed on top of the table (dismantled 19 inch monitor) and manipulated, informing the simulation. Agents adjust their routes through sneaky wayfinding algorithm ;-) Blue trails are highlighted based on occupancy per square meter in given time.
Hardware : PC + flat screen + PS3 Eye camera
Software : Processing + TUIO library
The approach of the project is to develop a theoretical framework and understand designing pervasive systems as an integral part of the urban landscape. The research was conducted in the city of Bath according to data provided by the CityWare project, in a heritage environment with tourist attractions.
It was proposed to investigate the application of research methods based on an agent system, employed by various practices in this field, as a framework for understanding the complex relationship between pervasive systems, and urban space in general. This project also attempted to propose a tool for the improvement of the existing infrastructure as well as a dynamic new transport planning.
The initial findings indicate that the pervasive system enables simulation the pedestrian flow in the city. The model is developed further in addition to generating the possibilities for the transport planning with the contribution of particle flow system and visual ability.