In this research, I try to programmatically generate building designs through infinite generative models of WFC algorithms and optimization of genetic algorithms. And focuses on the problem of slums in large cities in Shenzhen, China, Hong Kong, and the Bangladeshi community in London. In order to address these issues, I conceptualize an infinite generative vertical city that is primarily for education, but also accommodates other functions such as housing and recreational spaces. In addition, to make up for the shortcomings of the WFC algorithm in building optimization, I also employ genetic algorithms as a powerful tool for vertical urban environment optimization. By modeling the evolutionary process, genetic algorithms help to explore optimal building shapes and layouts while taking into account factors such as sunlight exposure, public space allocation, and occupancy density. In my research, genetic algorithms have been shown to help optimize building shapes and layouts to improve the environmental sustainability and livability of vertical cities. Through iterative improvement and selection, the genetic algorithm is able to identify solutions that strike a balance between building efficiency and environmental harmony.

In this research, my design process involves first optimizing the layout using genetic algorithms, and then adding architectural details with the WFC algorithm. The design rules for the entire building are derived from studying the design context. Due to the unique environment of slums, we need a vertical city for education. Additionally, to encourage residents to study hard through changes in the building’s environment, the building needs to interact with its inhabitants, controlling their behavior and allowing the building to rapidly adapt to new needs. This reflects the infinite possibilities of architecture, embodied in the concept of “endless.” This defines the design rules for the “endless vertical city in urban slum for education.” While algorithms have countless applications, the design rules based on the context are singular.