【作者】 梁成军；

【导师】 虞丽娟；

【作者基本信息】 上海体育学院 ， 运动人体科学， 2011， 博士

【摘要】 在竞技网球运动中技战术是影响比赛成绩的主要因素之一,运动员在比赛中技战术成功运用与否,往往会成为比赛胜负的决定因素。因此研究网球竞技比赛的技战术规律,为竞技网球运动日常技战术训练和比赛时技战术决策提供科学的支持,是提高我国网球竞技水平的一个重要途径。网球比赛中技战术复杂多变,对这类问题的准确科学决策往往很困难。信息技术的发展,使得决策可以在较多信息下进行,这也大大增加了决策的复杂性,使得依据个人的知识经验做出正确的决策受到制约。因此很有必要建立网球技战术决策支持系统,帮助人们探寻网球比赛技战术规律,辅助人们进行技战术决策,提高决策的科学性。本研究采用文献资料法、软件工程法、数据挖掘法和运动技战术建模研究法研究建立了网球技战术决策支持系统。系统以数据库、知识库、模型库三库结构为基础,采用VB6.0开发环境和SQL sever 2000数据库管理系统研究实现。论文研究了网球技战术决策知识结构体系和知识表示等相关知识库构建理论,构建与实现了网球技战术决策知识库;研究了数据挖掘理论与方法,实现了运用数据挖掘技术探求大量数据下隐含的技战术决策知识与规律;研究了系统动力学建模方法与原理,建立了可为技战术决策提供模拟决策支持服务的系列系统动力学模型。主要研究成果:(1)系统分析了网球技战术决策知识结构特征,建立了包括不同打法类型选手对抗决策、比赛中具体情况下技战术决策、不同比赛环境技战术决策三个大方面的知识库。知识的获取方法及来源主要是在专家的指导下对文献决策知识进行总结提炼和专家访谈的方式进行知识提取。采用产生式知识表示方法完成对知识的表示。同时建立了具有知识获取、编辑、检查和浏览等功能的知识库管理系统和基于知识推理的决策支持系统推理机。(2)研究了数据挖掘理论与方法,着重探讨了基于关联分析数据挖掘和基于粗糙集数据挖掘理论与技术两种方法在网球技战术决策领域的应用,并进行了具体的应用案例分析。关联分析可以发现隐藏于大型数据库中事物之间令人感兴趣的联系。研究采用Apriori算法建立了数据挖掘模型,实现了对比赛中技战术落点与得失分关联决策问题进行了挖掘分析。研究主要解决了最后两拍落点关联规则分析、发球轮前三拍落点关联规则分析、接发球前两拍落点关联分析等技战术落点决策问题。粗糙集是采用合适的算法对确定事物的多种属性进行约简,以去掉冗余属性,得到最小属性约简集;在此基础上对事物进行分类并生成决策规则,即得到知识。研究建立了粗糙集数据挖掘模型,并进行了数据挖掘的案例分析,解决了网球比赛中关键击球拍和连续几拍击球多属性决策问题。实例分析表明运用数据挖掘技术可以实现对击球落点、技术、线路、拍位等属性与得失分属性之间关系的数据挖掘分析,能够发现大量数据中隐含的技战术规律,弥补了统计分析的不足之处,完善了决策支持系统的功能。(3)通过技战术建模研究,开展比赛中各种技战术指标对比赛结果的影响的模拟与预测,也是技战术决策的一个方面需求。考虑到网球技战术决策需求特征,本研究采用了系统动力学建模方法理论,研究建立了系列网球比赛技战术系统动力学模型,为比赛的技战术决策提供模型模拟决策支持。研究建立了网球技战术主要统计指标的系统动力学模型、基于马尔可夫链随机过程的系统动力学模型和网球比赛技战术落点对抗模拟的系统动力学模型。改变运动员某技战术指标或因素的值,然后通过模型的模拟运行,可以找出影响运动员比赛得分率的主要技战术指标或因素,为技战术决策提供了科学有效的支持。本研究首次将决策支持系统理论应用到网球技战术决策领域中,依据决策支持系统、知识表示和数据挖掘等理论,针对网球教练员、运动员技战术决策的需求,建立了网球技战术决策支持系统,并进行了实际应用。更多还原

【Abstract】 In sports tennis,technique and tactics are one of the main factors that influence the match achievement, and it is also a decisive factor that determined the match result weather the athletes use it successfully or not in match. Therefore,it is an important way to improve tennis athletics level of our country to research on technique and tactics laws of the tennis match which aims to provide science technique and tactics decision making for tennis routine training and tennis match.Tennis technique and tactics are complex and mutable in tennis match, so it is difficult to make accurate decision-making for this kind of problem. The decision-making be made on the base of more information, and this will increase the complexity of decision-making and restrict accurate decision-making based on personal knowledge and experience. Therefore, it is very necessary to establish tennis technique and tactics decision support system, which can help people explore technique and tactics rules in tennis match, and aid people to make technique and tactics decision-making in order to increase the accuracy of decision-making.The tennis technique and tactics decision support system was built by adopting the methods of literature, software engineering, data mining and movement technique and tactics modeling. Based on database, knowledge base and model base, the system was realized on the base of VB6.0 development environment and SQL sever 2000 database. The knowledge architecture of tennis technique and tactics decision support and the theory of knowledge representation researched in this thesis, and tennis technique and tactics decision knowledge base was constructed and realized. Data mining theory and methods studied, and data mining technology used to explore technique and tactics decision knowledge and rules in large data. System dynamics modeling methods studied, and system dynamics models were built to provide simulation decision support. The research results of this thesis are as follows.(1) Based on Systematic analysis of the tennis knowledge structure character of technical and tactical decision-making, knowledge base was established that are including the technique and tactics decision-making of different play athlete antagonize, the technique and tactics decision-making of different game play situations, and the technique and tactics decision-making in different environment. Knowledge acquisition methods and sources include summarizing the literature under the guidance of experts and questionnaire. The production representation method was adopted to represent knowledge. This knowledge base management system,which consist of knowledge acquisition module, knowledge base editing module, knowledge check module, and knowledge base browsing module,and inference engine based on knowledge reasoning were established.(2) On the base of research theory and method of data mining, this research focus on the application research of the data mining theory and technology based on association analysis and rough set, and application case are provided.Association analysis can help people to find interesting, potentially relation between things from a large amount of database. The association data mining model was established based on Apriori algorithm,which was used to mine the association rules between stroke placement and stroke result (win or lose). The research solve such problems as the last two stroke placement association analysis, the first three stroke placement association analysis of the serve game,the first two stroke association analysis of receive game and other stroke placement decision-making problems.Rough set is adopting appropriate algorithm to reduce attributes that determine things in order to remove redundant attributes and get the least reduction of attributes set. On this basis, the things were classified and decision rules were generated, that is, for knowledge. The rough set data mining model was established,multiattribute decision-making problems were solved by using rough set data mining model,which include the crucial stroke and successive stroke decision-making in tennis match. The case analysis indicates that adopting this method inherence relation or rules between tennis stroke placement, technique, line, hand orientation and result can be capable of discovering, and this may make up shortcomings of statistical methods, and consummate the function of decision support system.(3)It is also a requirement of tennis technique and tactics decision-making that study on technique and tactics modeling and carry out simulation and prediction on how the indicators affect the game result. Considering the character of requirement of tennis technique and tactics decision-making, this research adopted system dynamic modeling method and established some system dynamic model of tennis technique and tactics, which can provide decision support on tennis technique and tactics in match with model simulation.Some system dynamic model were built in this research work including the main indices system dynamic model of tennis technique and tactics, system dynamic model of tennis technique and tactics based on Markov chain in stochastic process, and system dynamic model of tennis technique and tactics stroke placement simulation. When one of athlete′s indicator was changed, the simulation of the model can help people find the main index or factor that influence the match result, and this will benefit to technique and tactics decision-making.The decision support system firstly was applied to the field of tennis technique and tactics decision-making in this research.Based on decision support system theory, knowledge representation and data mining theory, for tennis coaches and players technical and tactical decision-making needs, the technique and tactics decision support system for tennis was established, and had been applied in practice.更多还原