Generative Approach For Architectural Layout Design Constrained By Topological Relationships

The topology of architecture is a widely-used model that specifies the adjacencies between rooms for architectural layout planning.It may be modified along with the modification of the design proposal during the design process.While most approaches regard it as an input constant and lack of the adjustment of topology according to the feedbacks from space.Approaches that generate high-quality results from partial and fuzzy defined topologies are rear.In this case,this research introduces two strategies for layout planning from topology:the first divides the production of topology and space as two separated parts.Plausible topologies are generated by computer program based on user specified objectives and are selected by designer for layout planning.The second generates both topology and space simultaneously.The topology is partial specified and the space is adjusted according to the missing information,such as the number of rooms or the adjacency relationships,to search for solutions that match the design objectives.The strategies are implemented using JAVA language and are tested in computer.The results imply that,firstly,partial-specified topology can lead to ideal results.Secondly,both strategies have vantages and disadvantages and are applicable for different applications.Thirdly,topology has limitations and thus is not applicable for all applications without involving or introducing extra models.