电力营销决策支持系统的研究和设计

【作者】 陈秀寓；

【导师】 王伟；

【作者基本信息】 大连理工大学 ， 控制理论与控制工程， 2003， 硕士

【摘要】 本文以电力营销系统为研究背景，针对电力企业改革后电力营销工作中一些急需解决的问题进行了研究。 电力营销系统与其他企业的营销系统一样，所面临的环境都是复杂多变的，因此有关电力营销的决策支持系统属于半结构化的决策支持系统范畴。这类系统中的决策含有大量不确定因素，缺乏程序化工作范式，需要意向决策支持的问题十分多见。本文提出了在电力营销管理系统中应用包含专家系统的推理模型思想，构建出了一个具有意向决策支持功能的电力营销管理系统框架，对如何建立问题生成子系统及其内部知识库进行了讨论。本文以一个具体问题为例，对其中模型库的构建进行了研究，建立了一种基于智能技术的数据挖掘模型，完成了电量销售情况的预测，为购售电工作提供了客观指导。本文提出了电力营销决策支持系统的总体设计方案，完成了其软件和硬件运行环境的设计。针对具体问题，给出了数据转换、多维预览、报表图表显示、MDX语句自引导、模型分析等多种功能的具体实现过程。 文中的电力营销系统经调试，可以正常、稳定的工作，基本上可以满足整个购售电工作的分析和决策的需要。对于提出的智能预测方法，进行了仿真实验，收到了较好的效果。对于所提出的在电力营销工作中应用意向决策支持技术的思想，给出了使用的具体例子。更多还原

【Abstract】 This paper takes power marketing system as working background, discusses some important problems coming from the reformation of electrical power corporation in our country.The environment which marketing system of electric power faces is very complex and changeable, just as the other marketing system. So decide support system (DSS) of power marketing belongs to the category of semi-structure. Decision making of this system has much uncertainty factor, lacks normal form to be followed, therefore many questions need intending decision support. Reasoning model based on expert system (ES) applying to power marketing system is described in this paper, meanwhile, the structure of knowledge base of this system is discussed in detail. Moreover the paper sets up a data mining model of on line analysis process database based on intelligent techniques, offers a method which is on the base of predecessors and combines ANN and fuzzy control to this problem, and designs the software system of power marketing DSS based on database warehouse running management for data transformation services (DTS), PivotTable service, multidimensional expressions (MDX) self-leading, model analyzing also. The paper focuses on intending decide support technique, data mining model and implementation of power marketing DSS.The power marketing decision support system has been tested running normally, and has reached the requirement basically. A simulation example of the buy-sell electric power forecasting shows better results.更多还原