Article:

The transport infrastructures are important elements in the cities, but, as there isa lack of planning, they tear through the urban fabric and leave empty spaces.Due to government and private disinterest, these spaces become vacant, forgottenand degraded. However, these extensive Terrain Vague offer new potential forurban use. To exploit this potential, we need methodologies that can offerpersonalised, extensive, feasible urban solutions. For this, we propose acomputational generative system, following a 4-step methodology: 1) Siteanalyses and Terrain Vague identification; 2) Site classification according toparameters based on a ``visual grammar''; 3) Algorithm associating spaceproperties with geometric transformation to generate solutions: namelytransformative operations in public spaces, additive transformations insemi-public spaces and subtractive operations in semi-private spaces; 4) Solutionevaluation and development, according to shade criteria, spatial hierarchy andvolumetric density. With our own algorithms combined with genetic algorithms,we guided the evolution of 50 volumetric solutions. The exponential increase ininformation requires new methodologies (Schwab, 2018). Results show thepotential of computational methodologies to produce extensive urban solutions.This research, developed in a final graduation project in Architecture, aims atstimulating generative methodologies in undergraduate courses.