Research And Application Of Knowledge Push Technology Oriented To Inteligent Design Of Product

Intelligent design of product is main content of the intelligent manufacturing,while knowledge push technology for product design is a key part of intelligent design.Design knowledge comes from the constant summary and discovery of designers in product development,and has important practical significance for product design and R&D activities.However,with the continuous expansion of design knowledge,the situation designers are faced with becomes very complicated in acquiring the knowledge they need from enormous knowledge which is of multi-source and heterogeneous but lack of organization.Therefore,to provide designers with personalized knowledge efficiently and actively has become a key issue to improve product design efficiency and shorten product development cycle.To solve above problems,this paper makes an intensive study from three aspects,namely,the representation of design knowledge,the descrption of designers' needs and the pushing algorithm of design knowledge.The main contents are as follows:In Chapter 1,the background and significance of the research in knowledge push technology are summarized.The reviews of the representation of design knowledge,the description of designers' needs and the pushing algorithm of design knowledge are given,and the main contents and framework of this paper are introduced.In Chapter 2,a knowledge representation method of product design based on design process is presented.The knowledge of product design is discussed.The theory of ontology,including concepts,language and building methods,are studied.The ontology of product design is built,based on which a knowledge model of product design is presented so that the knowledge is associated with the design process.The unified representation model for product design knowledge is established to obtain a unified expression for heterogeneous knowledge.In Chapter 3,multi dimension hierarchical knowledge requirement model based on users'behavior is presented.The concept of the model,the representation method and the characteristics of the users' behavior are discussed.The model includes three dimensions of design tasks,designers,and themes.The themes contains a number of thematic words that describe the user's needs and the measure value which is computed by GRA method.In Chapter 4,a hybrid collaborative filtering knowledge push algorithm by integrating items and users is presented.The three main components of neighborhood-based recommendation methods are discussed.The item-based CF and user-based CF are the introduced,and to improve the flaws of these algorithm,new algorithms based on content similarity and preface graph are presented.The similarity is computed through the knowledge model and requirement model built in former chapters.A hybrid collaborative filtering knowledge push algorithm is presented.MSD cost function and gradient descent method are used to obtain the parameters.The feasibility and effectiveness ofthe algorithm are verified by experiments.In Chapter 5,a knowledge push system for product design knowledge is developed.The framework of the system includes user interface layer,business logic layer,system service layer,model mapping layer and data resource layer.The function module of the system consists of knowledge reuse module,knowledge management module and knowledge push module.A case of shaft design is given to introduce the knowledge push module.In Chapter 6,the main content and innovation of this paper are summarized,and the future works are discussed.