Research On The Methods And Applications Of Extension Intelligent Design

Extenics is a new discipine that studies the extensive possibility of things, as well as the rules and methods of exploitation and innovation with formalized model. In this research, the limitations of current intelligent design are analyzed. Extenics is used as the means of knowledge engineering, to solve the problems of modeling, reasoning, and evaluation in diffenet phrases of current intelligent design, such as shceme deisgn, optimum design, and redesign.First of all, Extension intellignet design model (EIDM) is built base on Extencis. Definitions, representations, and intermediate codes for EIDM are present. Problem identification and integreated problem-solving method based on EIDM are discussed.To solve the limitations of current knowledge reasoning in intelligent design, rhombus thinking mode, a kind of creative thinking, is applied to the EIDM. The rhombus solving strategy is created. It is composed of matter-element extended reasoning and condition extension-set ceation, respectively implementing the divergence and the convergence of rhombus thinking. Extension constraint graph is proposed. The properties, the algorithms, the conflicts and the resolution of Extension constraint graph are present. Methods of matter-element extended reasoning based on extension constraint graph is discussed, including the extension inherit reasoning, the expansion reasoning, the correlative reasoning and the implication reasoning. Methods of calculating the depend function and creating the condition extension set in different cases are present as well.Case-based reasoning (CBR) is also an important part of shceme design. To slove the prolems in case representation and case retrieval, EIDM is applied in case-based design. Extension case-based reasoning (ECRM) is built. Distance, which is redefined in Extenics, is expanded to hyperspace and used to compute the similarity degree from existing cases to the object hyperspace. Thus, case retrieval from case point to interval space is implemented.Optimization modeling is always a difficulty in pratical optimal design. To solve this problem, theory of correlative network in Extenics is applied to create the intelligent optimization model. Optimization knowledge representation is firstly built. The correlative network is created based on the optimization knowledge representation. Based on correlative network, methods of generating three elements of the optimization model are proposed in detail. Methods of optimization calculation after modeling are also discussed.To solve the problems of lacking backtracking information in current sheme redesign, qualtitive reasoning and the correlative reasoing are combined to create the redesign solving strategies. Definitions of qualititive matter-element, matter-correlation network, qualtitive correlative networks (QCN), and theroms for qualitative correlative inferencing are present. On the basis of them, methods of generating redesign solving strategies are proposed. Object correlative matter-elements are created by building the matter-correlation network. By creating the object mechanisms and the qualitative correlative network, qualitative correlative inferencing is implemented, on which redesign solving strategies are generated.An intelligent design prototype is finally developed based on EIDM, called Extension Intelligent Design System (EIDS). Architecutreal structure, implementation mechanism, and functional modules are introduced. Knowledge reasoing, case-based reasoning, intelligent modeling in EIDS are respectively applied in parameter-evaluation, electromegnetism deisgn, and electromegnetism optimization modeling of the hydro-generator design in Zhangtan hydropower station. Results are satisfying, demostrating the validity and the practicability of the methods of Extension intelligent design...