Research On Ontology-based Case Representation And Retrieval For Design Of Building Structure

Building design is a typical knowledge-intensive complex product design which is a process including multi-level, multi-stage and multi-step. Designers often need to work with the aid of the knowledge and experience included in large number of existing design files. However it’s impossible for the designers to recall all the similar design files quickly and accurately, it’s necessary to adopt design archives management system to realize design knowledge reuse. But at present, there is no enough structured design knowledge management method for domestic building design companies to retrieve design files effectively, and gives rise to unnecessary repetitive design and poor efficiency for building design companies. Therefore, it’s imperative to change the traditional archives management mode and investigate the design knowledge reuse based on information technology under the digital and networked environment.Building design process involves a great deal of knowledge and experience, which are so complicated that can’t be represented with simple rules or models. Case-based reasoning (CBR) is a knowledge-based problem solving and learning method, and can reuse past experience and specific knowledge in the context of specific problem situations (cases) to solve new problems. CBR is especially applicable for the areas that are difficult to be represented with uniform rules or models and require rich experience. Therefore, it’s highly appropriate to adopt CBR to manage the design case knowledge. In addition, with the development of the Semantic Web, the advantage of applying ontology in the field of knowledge reuse and information retrieval is becoming increasingly apparent. Representing and organizing case knowledge by ontology can overcome the shortcomings of traditional CBR systems effectively.For these reasons, ontology is introduced into CBR method and case representation and retrieval method for design of building structures are studied in this paper. The major research work in this paper is as follows. Firstly, the background and significance of this study are elucidated, and the feasibility and importance of ontology-based CBR for design knowledge management are also analyzed. Secondly, the theoretical knowledge of CBR and ontology are introduced in detail to provide a theoretical basis for further study. Then building structure design ontology framework is constructed based on the existing ontology construction methods and ontology modelling is achieved in Protege platform. Finally, a method for the case representation and retrieval is presented. On the one hand, the case features and design instances are defined in formalization by ontological feature vector representation method. On the other hand, the strategy of two-level case retrieval is adopted, i.e., design cases are filtered out by means of SPARQL query sentence, and then the similarity of each attribute of data property and object property is calculated respectively. Analytic hierarchy process (AHP) is used to determine attribute weight, and the global similarity is calculation and ranking are achieved based on the nearest neighbor (NN) method to realize retrieval of design cases.