Searching spatio-temporal patterns in urban areas, using Artificial Neural Networks
Romera Giner, Juan Pedroa; Durán Fernández, Joséb
aDepartment of Architectural Projects, Universitat Politècnica de València, Spain bDepartment of Architectural Projects, Universitat Politècnica de València, Spain
Abstract:
The city must be understood as a complex system. A system with holistic properties where each of its elements and the relationships established between them form a whole. If we transform a part, the whole system is affected.
However, there are certain patterns that crystallize and repeat themselves over time. "Each pattern describes a problem that occurs again and again in our environment": Alexander (1977) and these patterns propose the core of the solution to the problem.
The objective of this research is to trace a methodological way to search for these urban patterns. The following work proposes a trip of short stops divided into five texts. In the first place, the analysis and simplification of the complex system of the city from urban models is approached. These models, developed computationally, will serve as an experimental laboratory and will allow us to establish possible urban predictions. Finally, the possibilities offered by the Artificial Neural Networks (ANN) for the search and detection of spatio-temporal patterns in the Modern City will be explored.
To conclude, this paper tries to establish a new methodological paradigm. Expanding urban designer tools to build the city as a whole: 'a system that generates systems'
Keywords: Complex-system; Spatio-temporal Patterns; Urban model; Artificial Neural Networks.