【作者】 蒋元涛；

【导师】 王宗军；

【作者基本信息】 华中科技大学 ， 管理科学与工程， 2004， 博士

【摘要】 随着加入WTO和市场化体系的逐步完善,我国的改革开放迈开了更大的步伐,各类企业都面临着巨大的机遇和挑战,如何把握机遇、迎接挑战是每个企业都必须认真思考的问题。然而,我们看到,在一些企业蓬勃发展的同时,也有很多企业在激烈的市场竞争面前束手无策、无能为力。计划经济模式下的经营方式,仍然束缚着企业适应市场经济的步伐,由于缺乏长远的战略方向,导致很多企业陷入严重的困境。近年来,随着现代科学技术突飞猛进的发展,计算机和互联网技术的不断突进和广泛应用,人类开始告别以资本、技术为特征的工业社会,正在进入一个崭新的时代。企业面对的战略环境更加恶劣,战略决策对企业变得比以往任何时候都重要。为了作出正确的战略决策,需要把专家的定性知识或定性判断同定量模型以及定量数据的描述有机地结合起来,以实现定性知识与定量数据之间的相互作用,仅仅靠传统的人工方法进行处理,依人的脑力去解决几乎是不可能的。因此,必须借助计算机寻找更加有效的战略决策方法,研究适合高层应用的辅助战略决策的计算机支持系统。七十年代出现了面向企业战略决策的执行支持系统和战略信息系统,而后发展到战略决策支持系统。作为决策支持系统(Decision Support System,简称DSS)研究的重要分支之一。SDSS的结构框架、设计原理在DSS的基础上向前发展,刻意强调面向企业高层决策者。本文首先根据决策组织的层次,应用从定性到定量综合集成的思想,提出了解决复杂战略问题的递阶分解思想。根据这个思想,探讨了定性模型和定量模型的集成模式和定性知识和定量数据之间的转换映射关系,并且提出了集成定性推理和定量计算的DSS体系结构和智能问题处理系统的分析处理流程,简要讨论了各功能模块。接着,在传统SWOT分析的基础上,把战略选择过程细化为满足两个约束条件地过程:战略地位和战略匹配点,从而构造出基于SWOT的双约束战略决策模型(SWOT-Based Double Retricting Strategic Decision Model,简称为SDRSDM)。利用SDRSDM,决策者首先确定出企业的战略地位,明确本企业在竞争环境中的战略位置,确定战略匹配点,从而客观、理性的在匹配战略中找到合适的战略方案。在以上分析的基础上,针对战<WP=4>略环境复杂、动态的特点,提出了SWOT分析的理性化思路,以投资企业现金流量灾变分析为例子,探讨了建立其它预测模型的可行处。然后,吸收定性和定量集成的思想,在分析影响因素的特性基础上,结合专家的定性判断能力和实际的定量计算模型,根据战略影响因素的定量、定性特征,引入更加合理的分值确定和权重赋值方法,以提高战略因素分析的可信度。在下面,通过引入有序决策树(Ordered Decision Tree, ODT)和最小有序决策树(The Smallest Ordered Decision Tree,SODT)概念将SDRSDM的求解进行了有利于建模和推理的形式化表示,从而将求解过程统一于整个ODT的不同节点之中。在此基础上,研究了SDRSDM的推理算法,为SDRSDM的系统实现奠定了基础。另外,通过以上分析,SDRSDM的求解是一个异常复杂的综合决策系统。因此,提出了基于人工神经网络的战略决策学习模型。研究了应用ANN对复杂的SDRSDM进行学习的原理。基于三层BP网络学习算法建造了评价专家用于决策分析的ANN学习模型,可以在实际应用中发挥很好的作用。最后是在前面各章工作的基础上,设计并实现一个基于SDRSDM的智能型动态战略决策支持系统应用原型,详细分析了系统的体系结构、功能特点、开发环境和流程以及关键的实现技术。并且为了适应战略决策的需要,提出了网络环境下集成系统的技术方法。更多还原

【Abstract】 As our country has jioned into WTO and the economic system has been perfected gradually, the reform and opening policy of China has maken great progress, all kinds of enterprises have to face the opportunity and challenge of the development, every one should think deeply about that how they can hold the chance. But, there are some companies which almostly have no ability to face the intense competition. In the operating mode of planned economy, many enterprises got into a hobble due to the enterprises’ culture of “iron bowl” and “mess”. Recently, with the development of scientific technology and wide application of computer and internet, human is stepping into a new age. The strategic situation that enterprises face, is bacoming badder. Strategic decision is more important than any time. Expert’s qualitative knowledge or qualitative judgments and quantitative model and quantitative datum are needed to be combined hormonally in order to realize interaction of qualitative knowledge and quantitative datum, but it is impossible only by method of traditional manpower and brain of human being. So more efficient strategic decision-making by computer should be utilized researching computer support system adapting to utilization of high level. In order to meet the need of strategy decision-making, in 1970’s executive system facing strategy decision-making of firm and stategy information system based on MIS and developed into SDSS. As one of research branches of Decision Support System (short for DSS), the framework of SDSS and design principium of SDSS developed forward focusing on that facing high decision-maker of firm.Firstly, this thesis applies the thought of integrating qualitative method and quantitative, bringing thought of disassemble by degrees and the integration of qualitative information and quantitative data. By this thought, discussing the integration mode of qualitative model and quantitative one, as well as the switching map relation between qualitative knowledge and quantitative data, bringing the DSS system structure of <WP=6>integrating qualitative reason and qualitative calculation and analysis flow of intelligent system of dealing with question. Every function module is discussed simply. Then, on the basis of traditional SWOT analysis, the strategy choice is divided into the satisfaction of two restricts: strategy status and strategy project. So, the SWOT-Based Double Restricting Strategic Decision Model, SDRSDM, is established. By SDRSDM, decision-maker ensures strategic status of enterprise firstly to show the strategic station of enterprise in competition environment. The suitable strategic project is found in matching strategy objectively and reasoningly. According to the complex and dynamic character of strategic situation, the reasoning thought of SWOT analysis is put forward. Making the risk analysis of investment enterprise as example, discussing the possibility of building other predict model. Then, absorbing the thought of integrating qualitative analysis and quantitative and based on the effecting factors analysis as well as the qualitative judgement ability of experts and quantitative calculating model, the reasonable mark and power confirming method is established to raise the reliability of factor analysis. As the following, the ODT and SODT are developed to express the solving process of SDRSDM, that benefit for the model and reason. So, the solving process is united in different nodes of the whole ODT. In this base, the reason is studied so that the base of SDRSDM is settled. In addition, we find that solving SDRSDM is a abnormal complex decision system. So, based ANN, complex strategic decision support system is developed. The ANN studying elements for complex SDRSDM is studied. Based the three-layer BP study arithmetic, the AMM model of studying the knowledge of experts is established. Finally, based on the analysis of above chapters, a apply prototype of intelligent dynamic strategic decision support system, that is based on SDRSDM, is designed and implemented. System’s basi更多还原