Abstract
This paper provides a comparative analysis of vector field methods for automatically generating urban fabric geometries and introduces the Rhino/Grasshopper plugin UrbanFab which implements these methods. Automated generation of urban design is a growing area of research addressing the complex challenge of building sustainable and economically viable cities. Within this research area vector fields are used to generate non-uniform urban fabric geometries. Similar to simulation of magnetic fields or moving fluids, vector field algorithms can be adapted to simulate urban sites, resolving complex site constraints and generating geometry used in the design of roads, land parcels or buildings. Vector field methods for urban fabric generation, however, are not well represented in computational tools serving urban designers, or in description of algorithms and evaluations in computational urban design literature. To address these challenges, this paper describes multiple urban vector field algorithms, with accompanying streamline visualization and evaluation methods. A comparative analysis of the results generated by these methods provides a means for designers to make informed decisions on which method is appropriate to their requirements and supports discussion of future work in urban fabric generation.