服务模式的发现与演化方法研究
发布时间：2018-10-13 13:38
【摘要】：随着互联网的高速发展,在大数据时代的大服务研究背景下,Web服务以更丰富方式呈现在日常生活中。单个服务之间通过合理组合形成大粒度服务方案,满足越来越多的用户个性化需求。一方面,海量的服务需求对服务组合的效率有非常高的要求,另一方面,用户对服务的期望不仅要求快速便捷,更希望获取最满意的服务结果。基于以上背景,服务模式这一概念被学者提出并引入到服务组合过程中,其核心思想是将高频使用的有效小粒度服务组合结果形成服务模式,以服务模式作为组合基础进行服务方案的快速高效构造,有效地提升组合效率和用户满意度。在服务模式的相关理论研究中,服务模式的设计、发现与演化技术是基础和核心工作。本文以服务模式的理论研究为背景,研究了服务模式的概念、描述、以及核心性质与使用方法,提出了服务模式的发现算法和演化方法,并在健康养老服务领域中进行了应用验证。在理论研究部分,首先研究了服务模式的概念及特性,给出了服务模式的形式化表达,对其核心的先验性、相似性和关联性三个性质进行了分析。然后,总结服务模式的特征,将服务模式抽象成图的形式,结合频繁子图挖掘技术提出了一种基于图挖掘的服务模式发现的算法。通过大量详尽的实验,评价了算法的性能,分析了算法各参数间的关系及对算法的影响。采用基于模式的收益最大化的服务组合算法来衡量所发现的服务模式的优劣,为算法的实际应用提供了指导。进而,研究了服务模式的演化规律与方法。本文为服务模式及单个服务制定了一套评估指标及其计算方法,分析发现了八种服务模式的演化类型,并给出了针对各类型变化的模式演化策略,使得挖掘的服务模式不断更新优化,处于最大程度的可利用状态。在应用验证部分,将提出的各算法和技术应用到系统设计中。设计实现了服务模式库管理工具,对服务模式进行有效的发现、管理、演化更新等操作。对所提的发现方法进行了应用验证,同时在健康养老服务领域中进行了应用实施,验证了所提方法的正确性和有效性。
[Abstract]:With the rapid development of the Internet, Web services are presented in daily life in a richer way under the background of the research on the big services in big data era. A large granularity service scheme is formed by rational composition between individual services to meet more and more users' personalized needs. On the one hand, massive service demand has a very high demand for the efficiency of service composition. On the other hand, users' expectations for services not only require rapid and convenient, but also hope to obtain the most satisfactory service results. Based on the above background, the concept of service pattern is put forward by scholars and introduced into the process of service composition. The service pattern is used as the composition basis to construct the service scheme quickly and efficiently, which can effectively improve the composition efficiency and user satisfaction. In the theoretical research of service pattern, the design, discovery and evolution of service pattern are the foundation and core work. In this paper, based on the theoretical research of service pattern, the concept, description, core properties and usage method of service pattern are studied, and the discovery algorithm and evolution method of service pattern are proposed. And the application in the field of health care for the aged is verified. In the theoretical research part, firstly, the concept and characteristics of service pattern are studied, the formal representation of service pattern is given, and the three properties of its core priori, similarity and relevance are analyzed. Then, we summarize the characteristics of service patterns, abstract service patterns into graphs, and propose an algorithm of service pattern discovery based on graph mining combined with frequent subgraph mining technology. Through a large number of detailed experiments, the performance of the algorithm is evaluated, and the relationship between the parameters of the algorithm and its influence on the algorithm are analyzed. The service composition algorithm based on pattern revenue maximization is used to evaluate the advantages and disadvantages of the discovered service pattern, which provides guidance for the practical application of the algorithm. Furthermore, the evolution law and method of service mode are studied. In this paper, a set of evaluation indexes and their calculation methods for service mode and single service are developed, and the evolution strategies for each type of service model are given, and the evolution types of eight service modes are found. The service pattern of mining is updated and optimized to the maximum extent. In the part of application verification, the proposed algorithms and techniques are applied to the system design. The management tool of service pattern library is designed and implemented, and the operation of service pattern discovery, management, evolution update and so on is carried out effectively. At the same time, the method is applied and implemented in the field of health care for the aged, which verifies the correctness and effectiveness of the proposed method.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP393.09
本文编号：2268841
[Abstract]:With the rapid development of the Internet, Web services are presented in daily life in a richer way under the background of the research on the big services in big data era. A large granularity service scheme is formed by rational composition between individual services to meet more and more users' personalized needs. On the one hand, massive service demand has a very high demand for the efficiency of service composition. On the other hand, users' expectations for services not only require rapid and convenient, but also hope to obtain the most satisfactory service results. Based on the above background, the concept of service pattern is put forward by scholars and introduced into the process of service composition. The service pattern is used as the composition basis to construct the service scheme quickly and efficiently, which can effectively improve the composition efficiency and user satisfaction. In the theoretical research of service pattern, the design, discovery and evolution of service pattern are the foundation and core work. In this paper, based on the theoretical research of service pattern, the concept, description, core properties and usage method of service pattern are studied, and the discovery algorithm and evolution method of service pattern are proposed. And the application in the field of health care for the aged is verified. In the theoretical research part, firstly, the concept and characteristics of service pattern are studied, the formal representation of service pattern is given, and the three properties of its core priori, similarity and relevance are analyzed. Then, we summarize the characteristics of service patterns, abstract service patterns into graphs, and propose an algorithm of service pattern discovery based on graph mining combined with frequent subgraph mining technology. Through a large number of detailed experiments, the performance of the algorithm is evaluated, and the relationship between the parameters of the algorithm and its influence on the algorithm are analyzed. The service composition algorithm based on pattern revenue maximization is used to evaluate the advantages and disadvantages of the discovered service pattern, which provides guidance for the practical application of the algorithm. Furthermore, the evolution law and method of service mode are studied. In this paper, a set of evaluation indexes and their calculation methods for service mode and single service are developed, and the evolution strategies for each type of service model are given, and the evolution types of eight service modes are found. The service pattern of mining is updated and optimized to the maximum extent. In the part of application verification, the proposed algorithms and techniques are applied to the system design. The management tool of service pattern library is designed and implemented, and the operation of service pattern discovery, management, evolution update and so on is carried out effectively. At the same time, the method is applied and implemented in the field of health care for the aged, which verifies the correctness and effectiveness of the proposed method.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP393.09
【参考文献】
相关期刊论文 前10条
1 庄亚欧;徐汉川;徐晓飞;王忠杰;何霆;;基于服务模式的快速服务组合方法研究[J];计算机工程与应用;2016年01期
2 石杰;;基于经典Apriori算法的频繁模式挖掘算法[J];电子技术与软件工程;2015年07期
3 李亮;陈莉;李华;王珊珊;张敏超;;一种改进的频繁子图挖掘算法[J];计算机与应用化学;2014年02期
4 郭春丽;李海波;;基于可变粒度的服务工作流系统服务质量可控性研究[J];中国机械工程;2011年21期
5 王爱平;王占凤;陶嗣干;燕飞飞;;数据挖掘中常用关联规则挖掘算法[J];计算机技术与发展;2010年04期
6 唐德权;朱林立;;频繁子图挖掘算法研究[J];计算机工程;2009年09期
7 吴甲;陈];;一种快速的频繁子图挖掘算法[J];计算机应用;2008年10期
8 李先通;李建中;高宏;;一种高效频繁子图挖掘算法[J];软件学报;2007年10期
9 胡海涛;林碧英;;一种基于组合模板的大粒度服务组合方法[J];华北电力大学学报;2006年05期
10 王艳辉;吴斌;王柏;;频繁子图挖掘算法综述[J];计算机科学;2005年10期
相关博士学位论文 前1条
1 王显志;大规模需求环境下基于服务模式的服务组合优化方法[D];哈尔滨工业大学;2014年
相关硕士学位论文 前1条
1 刘振;频繁子图挖掘算法的研究与应用[D];中南大学;2009年
，本文编号：2268841
本文链接：https://www.wllwen.com/guanlilunwen/ydhl/2268841.html
