Research On Design And Reuse Extension Strategy Generating System Framework

With the purpose of improving the deciders' decision-making and the level of machine intelligence, Extenics proposes to research and design Extension Strategy Generating System (ESGS) which professionally solves the incompatible problems. The main idea of designing ESGS is, describing the problem's conditions and objectives with formalized Extenics Language, modeling for the core problems, analyzing core problems with rhombus thinking method, carrying out homologous transformation, so as to generate some strategies which meet or close to the user's needs.However, it's not an easy thing for developers to develop an effective ESGS, especially to those who don't have the knowledge of Extensics. If we can design and develop the Extension Strategy Generating System Framework (ESGS-F), reuse the design and code of mature strategy generating system sufficiently, we can greatly reduces the the difficulty of system development and improves ESGS's development efficiency.Both at home and abroad, even the development of ESGS is still in the exploration stage, not to mention the ESGS-F. Therefore, with the basic theory of design pattern, interface-oriented programming and Extenics, this thesis tries to design and achieve the general framework of generating strategies for incompatible problem. During this process, general architecture, controller, general components and its'cooperation have been mainly explored.In the process of developing framework, multiple iterative methods were used, so that the ESGS-F's construction and ESGS's construction promoted each other. On the one hand, after collecting and analyzing the typical application in specific areas, then getting domain model of common areas (which can share components) and Well-defined variability (extension point), we could design and implement the ESGS-F, which finally used in Travel-ESGS and Tenement-ESGS. On the other hand, the feedback ESGS-F problems in the application development process further improved ESGS-F.This project has designed and implemented the ESGS-F finally. What's more, the ESGS-F's stability, rationality, understandability and reusability have been confirmed by two ESGSs.The results show that the design and code of ESGS-F can be generally applied to the incompatibility of the strategy generating system development; the well-defined extension points of ESGS-F can be expanded accurately in specific areas of strategy generating system; ESGS-F using could be more efficiently develop a viable and stable strategy generating system.The process of exploring this subject, is a bold attempt from scratch, and has filled the gaps in the research of ESGS. Also, it has high research and application value. Its innovation lies in the following two aspects:(1) Different from the general problem, this article discusses strategies generating mechanism which solving contradictions specifically. And, based on this mechanism, corresponding intelligent strategy generating system was developed. It basically achieved those research objectives:finding the contradictions solving method; improving the deciders' decision-making and the level of machine intelligence.(2) In this thesis, after exploring the general laws of such strategy generating systems sufficiently, generaton system's reusable design and code which constitute ESGS's framework were abstracted. All these jobs will assist developers to build a stable and feasible strategy generating system for contradictions of specific areas efficiently.However, the ESGS-F current implemented is just more preliminary, pending further refinement.This work is supported by the Guangdong Provincial Natural Science Foundation (grant no.8151009001000041)—"Research on the Strategy Generating Mechanism of Software Agents" and Guangdong Provincial Natural Science Foundation (Grant no. 10151009001000044)—"The Client Values Research based on Extension Data Mining".