Research On Knowledge Acquisition And Evaluation For Biologically Inspired Design

The collaborative utilization of multidisciplinary knowledge is the mainstream of modern product development innovation.As the largest knowledge repository for human society,nature has optimized and generated numerous remarkable biological skills among the long period of evolution,which is an inexhaustible knowledge source for product development innovation inspiration.Biologically inspired design(BID)is a knowledge driven design paradigm,which aims to establish a systematic interdisciplinary knowledge analogy and to provide inspirations and enlightenments for engineering problem-solving and technical innovation through natural phenomena studies or biological systems imitation.However,the acquisition and evaluation of biological knowledge have been put a dilemma and puzzled the biologically inspired design for a long period,which seriously hindered the systematic application of biological knowledge.To tackle those issues,a systematic biological knowledge acquisition and evaluation methodology is proposed in view of relevant techniques in knowledge-based engineering.Furthermore,a computer-aided design prototype is developed and the validity is verified through some case studies.The main contributions of this dissertation are summarized as follows:(1)The natural language processing based biological case acquisition is studied.After an in-depth view on biological case acquisition,this dissertation proposes a text mining based techinique for biological case acquisition,and develops a scalable biological case base.In view of the vector space model and similarity measurement of natural language,a vector space model based biological corpus retrieval method is presented.(2)The biological knowledge modeling method based on functional case knowledge model is investigated.Based on the review of current biological knowledge representation models,this dissertation proposes a functional case knowledge model for biological case modeling.Moreover,a biologically inspired design based ontology association model is presented and a biological case base is constructed.The proposed knowledge model not only combines with functional modeling and case modeling,but also contains the influence of environmental factors.(3)The feature selection method and knowledge clustering techniques for biological knowledge model are studied.Based on the attribute significance of knowledge model,this dissertation presents an attribute reduction method and conducts the feature selection for biological knowledge model.Besides,the clustering structure construction method for biological case base is studied.The lateral clustering is constructed,the longitudinal growth of the clustering structure is presented,and the growing hierarchical structure model of case clustering is developed.(4)Research on biological principle solution evaluation under subjective and uncertain environments is conducted.For the biological principle solutions derived from text mining and knowledge clustering,this dissertation establishes the evaluation criteria and presents an uncertainty representation model for subjective information under group decisionmaking environments.Furthermore,the framework for biological principle solution evaluation is developed.The criteria weighting and alternative ranking methods with subjective and uncertain informations are proposed.The subjectivity and uncertainty measurement for group decision-making environments is investigated.(5)The prototype is developed and case studies are presented.Based on above techniques,this dissertation develops a computer-aided prototype for biologically inspired design.Then the conceptual design of industrial products such as multiple parallel-flow separatory microchannel heat exchanger and oblique multi-petal envelope check valve are taken as examples and the validity of the proposed method are verified.