【作者】 陈文；

【导师】 杨善林；

【作者基本信息】 合肥工业大学 ， 管理科学与工程， 2007， 硕士

【摘要】 面对汽车消费市场中日益剧增的竞争压力，不断推出新车型已成为汽车厂家争夺市场份额的重要手段，这就对传统的车型决策过程提出了挑战。本文以如何科学、快速地进行车型决策为出发点，分析了传统车型决策过程中存在的问题，将先进的Internet技术和群决策科学理论应用到车型决策过程中，设计开发出基于Internet的车型决策平台，并验证了该平台的有效性。论文主要研究工作和创新性成果有以下几个方面：(1)在对多家轿车生产厂家进行详细调研的基础上，分析了影响轿车造型设计的专业领域，建立了系统的车型决策指标体系，并对确立的指标体系进行了详尽的解析。(2)针对目前轿车车型决策过程中存在的问题，将先进的Internet技术和群决策等相关科学理论应用到车型决策过程中，比较选择了合适的决策方式和群决策方法，设计出Internet环境下的车型决策流程。(3)在给出向量相似度定义的基础上提出了一种新的群体一致性算法。通过计算专家判断向量间的相似度可准确地衡量专家的一致性程度，运用方差的大小来衡量群体一致性程度能快速找到存在最大评判差异的方案，本文给出的群体一致性算法较以往的群体一致性算法能加快专家群体思维收敛的速度，最终达到群体意见一致。(4)设计实现了基于Internet的车型决策平台，并验证了平台的有效性。该平台由发言讨论模块、专家评判模块、查询模块、群体一致性判断模块、集结模块以及后台管理模块等六大模块组成，运用该决策平台可以实现异地决策，从而为快速、科学地进行车型决策提供了保证。更多还原

【Abstract】 Due to the increasing competition pressure in car market, it is being a vital meaning for car manufacturers to constantly launch new car style to possess more market shares. That is a challenge for the process of traditional car styling decision. Starting with how to make car styling decision scientifically and quickly, this dissertation analyzes the problems in traditional car styling decision, implements both the advanced internet-based technology and group decision making, designs and develops an internet-based platform for car styling decision and finally testifies the effectiveness of the platform.The main creative works and achievements in this dissertation are as follows:(1) On the basis of detail investigation on many car manufacturers, this dissertation analyzes the professional fields influencing the car styling, valuates an index system and interprets it specifically.(2) Aiming at the problems existing in the process of car styling decision, this dissertation introduces advanced Internet technology and group decision making, compares and chooses a better decision-making method and designs the process of car styling decision under the Internet Environment.(3) Based on the definition of vector similarity, I suggest a new Group Consistency algorithm which could accurately assess the experts consistency by calculating the vector similarity judged by the experts and could quickly find the program which has the biggest evaluation difference by using the variance calculating the Group Consistency. Compared to the previous one, the new Group Consistency algorithm could speed up the convergence of expert group thought and ultimately reach the agreement.(4) This dissertation designs and develops an Internet-based platform for Car Styling Decision and verifies the effectiveness of it. This platform has six components: Speaking and Discussing module, expert evaluation module, Query Module, group consistency judgment module, assemblies module, as well as the background management module. This platform could also be used to achieve the remote decision-making which could provide a guarantee for the quick and scientific Car Styling Decision.更多还原