A Self-Organizing Neural System for Urban Design (2001)
article⁄A Self-Organizing Neural System for Urban Design (2001)
abstract⁄The focus of this research is the development of an urban simulation system and its use to analyze growth factors in an urban design proposal. Unlike predictive simulation models, which attempt to accurately simulate future conditions resulting from a proposal, our neural network model is tuned to creatively present socioeconomic deficiencies and requirements for proposed developments. The system is built using a novel variant of Kohonen’s selforganizing neural map algorithm. Urban data of a simulated region is embedded in the neural net and correlated, in varying degrees, with data obtained from case study andor other local regions. By projecting design ideas onto this network, designers gain an insight into the proposal’s impact based on complex, nonlinear relationships of socioeconomic data, which are otherwise difficult to envision.
