A Preliminary Study On Data Driven Architecture Intelligent Design Program

In the age of information,the ability to analyzing and manipulating data in many the industrial circle has reached an unprecedented height.The architectural industry now also has an unprecedented trend to embrace cutting-edge computer technologies.Within a long-term development of architecture design industry,a huge amount of data has been accumulated.These data can be described as big data.By analyzing,producing and proper using of big data,many hard problems considered in the past now can be addressed.By using architectural big data and data-driven method,architecture intelligent design can be hopefully achieved.On the basis of previously mentioned background,this thesis took a preliminary study on architecture intelligent generation program on single-loaded small-sized apartment.The major research work as follows:1.By analyzing some typical design theory and practice on modern architecture history,the thesis founded the concept of architectural data,then analyzed the form,characteristics and usage of architectural data.Next,the thesis proposed a concept of architectural big data,and revealed that architectural big data can be accumulated naturally,then discussed data-driven method as a proper way to utilize architecture data.Meanwhile,the thesis discussed that we can make most use of big data by using data driven method,and finally discussed the characteristics and application methods of today's data driven method.2.By analyzing the development of half century's development of computer-aided design history and automatic design thinking and the intelligent application implementation,we summarized the characteristics,limitation,and breakthrough of traditional intelligent design methods,and revealed the characteristics and core principles of data driven intelligent design application.3.The thesis proposed a kind of data driven intelligent design program on small-sized apartment.Firstly,the thesis analyzed current study of architectural plan database and bigdata technology,demonstrated that we can collecting and analyzing big data.Secondly,on the assumed basis of gathered big data,we proposed a data-driven,architect-friendly architecture functional space generation program,and implemented a prototype to verify the feasibility of the proposal.The first part of the program can provide a service to query plans by architectural features and other metadata,and play a role at design analysis stage.The second part is and data-driven plan generation program,it can analyze inputted design requirements and generate a plan state by querying the database for optimization.Then,by computing the cost function that indicates “deficiency score” of the current plan state,program use a stochastic optimization method to reduce function cost and finally generate a proper plan.The plan can be used as a reference information for architect when at design synthesizing stage.program performance is restricted due to the limited research time and resource,so it can only deal with single-loaded small-sized apartment.but with the accumulation of data and time,it can hopefully generate more proper and complex plan.