Abstract by Edward Doyle
A Computationally Creative Approach to Urban Design
As the population continues to grow in cities around the world, experts in the field of urban design continue to seek novel ways of designing cities that attract tourists, new residents, and businesses. This paper investigates the application of an artificial intelligence system with computational creativity (CC) emphasis to replace this process which has traditionally been performed by human designers. We collected geographical information for a set of popular cities and stored it in a database to use as a knowledge base for our agent. Using a set of matrices to represent the collected geographical information of different cities as a knowledge base, we implemented a modified genetic algorithm to generate designs for cities. The utility of these created cities are then judged using software that was developed by MIT's urban research lab, comparing them to the utility of the original cities.