Project description:

The Inner Jewish district was once the most vivid neighbourhood of Budapest. It has now completely lost its character of being the commercial heart of the city that it was 150 to 100 years ago. This district is unique in the city because it evolved naturally without any master-plans, completely driven only by market forces from the beginning of the eighteenth till the end of the nineteenth century. It was an organic adaptive system just like any system in nature adjusting itself to the changes of the environment. In the twentieth century short-sighted master-plans, as well as the economic crisis of the wars forced it into deterioration. Another unique feature of the site was its connectivity: the street pattern emerged in a completely evolutionary manner and passages through the inner yards appeared to provide further accessibility. The loss of commercial activities during the twentieth century forced those passages to get closed.

Two interconnected computer programs were developed in order to explore the possibilities to re-establish the unique qualities of adaptivity and connectivity of this district: a pedestrian movement simulation and a genetic algorithm.

Research on pedestrians’ behaviour provides data about basic rules that people follow when moving through space. These rules are different for people who know the place (locals) and people who explore it (tourists), and they are also vary for different times of the day or the night. The existing attractors (ground-floor retails) on the site are mapped, and the artificial characters are inserted into the digital environment from important nodes such as transportation terminals or busy pedestrian areas around the site. Guided by their basic artificial intelligence, the characters explore the site and find their targets. The paths that they choose with their movement are recorded and evaluated according to how intensely they are chosen. The busiest paths actually become attractors for new retail spaces to emerge on the site.

A genetic algorithm is then utilized to generate the distribution of new retail spaces emerging in the district. This is a simple, but effective computer program that performs a search in a ‘population’ of possible solutions, selecting the ones that most closely satisfy the criteria that are defined by the problem. Furthermore, the software mixes the features of the selected solutions, in order to create a new population (called offspring) of better solutions to the problem. The same process is repeated now within the ‘offspring’ population until an optimal solution to the problem is reached. This is how the computer is used to simulate the same evolutionary processes that once shaped the site, in order to create formations that would be as adaptive to the changing environment as the previously existing, naturally developed district.

The new distribution of spaces on ground-floor level means new attractors for pedestrians. This leads to a change of the chosen paths and the intensity of movement on them. The new data about pedestrian movement is fed back into the genetic algorithm which generates a new optimized distribution of retail spaces. This feedback process of ‘dialogue’ between the two programs is repeated until a favourable agreement between the criteria for adaptivity and connectivity of the pattern of urban spaces is reached.

Project title: URBANOIDS
Subtitle: Revitalizing the Inner Jewish District of Budapest: Adaptivity and Connectivity
Location: Budapest, Hungary
Authors: Fabiano Friedrich e Silva, Krassimir Krastev
Tutor: Neil Leach, Andrea Haase
University: Hochschule Anhalt, Dessau Institute of Architecture, Germany
Collaboration : Cornel Prahovean
Year: 2006