Integrating urban analysis, generative design, and evolutionary optimization for solving urban design problems
Reinhard Koenig,Yufan Miao,Anna Aichinger,Katja Knecht,Kateryna Konieva
DOI: https://doi.org/10.1177/2399808319894986
2019-12-18
Abstract:To better support urban designers in planning sustainable, resilient, and livable urban environments, new methods and tools are needed. A variety of computational approaches have been proposed, including different forms of spatial analysis to evaluate the performance of design proposals, or the automated generation of urban design proposals based on specific parameters. However, most of these propositions have produced separate tools and disconnected workflows. In the context of urban design optimization procedures, one of the main challenges of integrating urban analytics and generative methods is a suitable computational representation of the urban design problem. To overcome this difficulty, we present a holistic data representation for urban fabrics, including the layout of street networks, parcels, and buildings, which can be used efficiently with evolutionary optimization algorithms. We demonstrate the use of the data structure implemented for the software Grasshopper for Rhino3D as part of a flexible, modular, and extensible optimization system that can be used for a variety of urban design problems and is able to reconcile potentially contradicting design goals in a semi-automated design process. The proposed optimization system aims to assist a designer by populating the design space with options for more detailed exploration. We demonstrate the functionality of our system using the example of an urban master-design project for the city of Weimar.
environmental studies,urban studies,regional & urban planning,geography