The machine learning process in applying spatial relations of residential plans based on samples and adjacency matrix

The current world is moving towards the development of hardware or software presence of artificial intelligence in all fields of human work, and architecture is no exception. Now this research seeks to present a theoretical and practical model of intuitive design intelligence that shows the problem of learning layout and spatial relationships to artificial intelligence algorithms; Therefore, the question of this research is to present a theoretical and practical model through the integration of computational methods for machine learning to apply spatial relationships in residential plans based on architectural criteria and standards. The problem of this research is to find the precise methods and techniques of applying spatial relationships to architectural plans through artificial intelligence algorithms based on architectural criteria, which should be applied by automatic design intelligence in machine perception and then in the design process. Also, the hypothesis of the research shows that to achieve this, it is necessary to use adjacency matrices, calculation formulas, and various algorithms for the perception and production of spatial relationships; Therefore, the research method is explanatory research to examine and transfer complex ideas and information based on data collection, including library studies, documents, educational samples, calculations, and mathematical formulas. The explanatory method seeks to present new knowledge, describe a process, or develop a concept. As the results of the research show, instead of randomly starting plan generation through evolutionary algorithms, spatial relations in architectural plans can be done faster and more accurately by combining adjacency matrices, and computational formulas with evolutionary algorithms.