基于群智能优化的云制造资源服务搜索方法研究
发布时间：2019-01-10 15:27
【摘要】：目前，我国制造业正经历由生产型主导向服务型转轨的过程，实现制造业的全面服务化成为新的经济增加热点，然而要实现由传统生产模式转型发展成技术服务型为主导的生产模式需要相关的先进技术、崭新的服务平台以及全新模式的作支撑。随着云计算，物联网等新一代IT技术蓬勃发展，一些先进的制造模式和技术与其相融合形成了现在的云制造模式。云制造旨在汇聚全球资源，实现资源优化配置，为加快经济转型和传统制造业的升级提供了核心动力。 服务搜索作为云制造的核心技术之一是云制造实现资源优化配置和按需使用的关键使能技术。服务的搜索与发现为实现服务的组合与优选提供数据支撑，其性能好坏直接影响资源服务的优化配置。而现有服务搜索的研究大多数是针对计算资源和Web服务，在制造领域关于服务搜索的研究不多。由于制造资源的异构性、复杂性，导致不能直接应用现有的一些服务搜索方法。因此，本文将以Web服务发现研究方法为基础，结合制造领域资源的特点，对云制造服务搜索做出了深入的研究，提出了制造资源搜索整体系统框架，在制造资源虚拟化和服务化的基础上建立了基于QoS的服务搜索模型，并研究了基于蜜蜂算法等群智能优化算法的服务搜索方法。 首先，，在研究现有的服务组合与优选体系、关键技术和服务流程的基础上，提出了基于QoS服务搜索的系统架构，并且给出了实现基于QoS服务搜索的三个环节，即服务QoS的提取、服务QoS的评估以及服务QoS的比较。该体系给出了实现资源准确发现的技术路线和工作流程，为制造云服务匹配提供完整的信息支撑。 然后，针对现有的服务描述在描述制造资源服务信息不全面，分类不清晰的问题，着重分析云制造资源有的特征，并且用本体建模语言描述了制造资源的基本信息、制造功能信息、QoS信息。按照用户对服务性价比的要求，给出了QoS评价指标及计算公式并且建立一个四元组的服务发现模型和约束指标。 由于蜜蜂算法在收敛性和稳定性上都要优于传统的群智能优化算法，并能更快的发现复杂优化问题的最优解，并且具有较强的鲁棒性、易于与其他方法结合、优良的分布式算机制等优点。本文将增强型蜜蜂算法作为云制造资源服务搜索的算法。 最后，搭建了基于QoS信息云服务发现仿真测试环境。用改进后的蜜蜂算法对其进行测试与性能比较，对比实验表明，本文的搜索算法在收敛时间和收敛精度上更为快速和稳定。
[Abstract]:At present, China's manufacturing industry is undergoing a process of transition from production-oriented to service-oriented, and the realization of full-scale service-oriented manufacturing has become a new hot point in economic growth. However, in order to realize the transformation from traditional production mode to technology-service-oriented production mode, the related advanced technology, new service platform and new mode are needed. With the rapid development of cloud computing, Internet of things and other new generation of IT technology, some advanced manufacturing models and technologies are combined to form the cloud manufacturing model. Cloud manufacturing aims to pool global resources and achieve optimal allocation of resources, which provides the core power for accelerating economic transformation and upgrading traditional manufacturing industries. As one of the core technologies of cloud manufacturing, service search is a key enabling technology for resource optimization and on-demand deployment in cloud manufacturing. The search and discovery of services provides data support for service composition and optimal selection, and its performance directly affects the optimal configuration of resource services. However, most of the existing research on service search is focused on computing resources and Web services, but there are few researches on service search in manufacturing field. Due to the heterogeneity and complexity of manufacturing resources, some existing service search methods can not be directly applied. Therefore, based on the research method of Web service discovery and the characteristics of manufacturing resources, this paper makes a deep research on cloud manufacturing service search, and puts forward the overall system framework of manufacturing resource search. Based on the virtualization and service of manufacturing resources, the service search model based on QoS is established, and the service search method based on swarm intelligence optimization algorithm such as bee algorithm is studied. First of all, on the basis of studying the existing service composition and optimization system, key technology and service flow, the system architecture based on QoS service search is put forward, and the three links to realize QoS service search are given, that is, the extraction of service QoS. Service QoS evaluation and service QoS comparison. The system provides the technical route and workflow to realize the accurate discovery of resources, and provides complete information support for manufacturing cloud service matching. Then, aiming at the problem that the service description is not comprehensive and the classification is not clear, the characteristics of cloud manufacturing resources are analyzed, and the basic information of manufacturing resources is described by ontology modeling language. Manufacturing function information, QoS information. According to the requirement of service performance to price ratio, the QoS evaluation index and calculation formula are given, and a quaternion service discovery model and constraint index are established. Because the bee algorithm is superior to the traditional swarm intelligence optimization algorithm in convergence and stability, it can find the optimal solution of the complex optimization problem more quickly, and has strong robustness, which is easy to combine with other methods. Excellent distributed computing mechanism and other advantages. In this paper, the enhanced honeybee algorithm is used as a search algorithm for cloud manufacturing resource services. Finally, a simulation test environment based on QoS information cloud service discovery is built. The improved honeybee algorithm is used to test and compare its performance. The comparison experiment shows that the search algorithm in this paper is faster and more stable in the convergence time and accuracy.
【学位授予单位】：武汉理工大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.09
[Abstract]:At present, China's manufacturing industry is undergoing a process of transition from production-oriented to service-oriented, and the realization of full-scale service-oriented manufacturing has become a new hot point in economic growth. However, in order to realize the transformation from traditional production mode to technology-service-oriented production mode, the related advanced technology, new service platform and new mode are needed. With the rapid development of cloud computing, Internet of things and other new generation of IT technology, some advanced manufacturing models and technologies are combined to form the cloud manufacturing model. Cloud manufacturing aims to pool global resources and achieve optimal allocation of resources, which provides the core power for accelerating economic transformation and upgrading traditional manufacturing industries. As one of the core technologies of cloud manufacturing, service search is a key enabling technology for resource optimization and on-demand deployment in cloud manufacturing. The search and discovery of services provides data support for service composition and optimal selection, and its performance directly affects the optimal configuration of resource services. However, most of the existing research on service search is focused on computing resources and Web services, but there are few researches on service search in manufacturing field. Due to the heterogeneity and complexity of manufacturing resources, some existing service search methods can not be directly applied. Therefore, based on the research method of Web service discovery and the characteristics of manufacturing resources, this paper makes a deep research on cloud manufacturing service search, and puts forward the overall system framework of manufacturing resource search. Based on the virtualization and service of manufacturing resources, the service search model based on QoS is established, and the service search method based on swarm intelligence optimization algorithm such as bee algorithm is studied. First of all, on the basis of studying the existing service composition and optimization system, key technology and service flow, the system architecture based on QoS service search is put forward, and the three links to realize QoS service search are given, that is, the extraction of service QoS. Service QoS evaluation and service QoS comparison. The system provides the technical route and workflow to realize the accurate discovery of resources, and provides complete information support for manufacturing cloud service matching. Then, aiming at the problem that the service description is not comprehensive and the classification is not clear, the characteristics of cloud manufacturing resources are analyzed, and the basic information of manufacturing resources is described by ontology modeling language. Manufacturing function information, QoS information. According to the requirement of service performance to price ratio, the QoS evaluation index and calculation formula are given, and a quaternion service discovery model and constraint index are established. Because the bee algorithm is superior to the traditional swarm intelligence optimization algorithm in convergence and stability, it can find the optimal solution of the complex optimization problem more quickly, and has strong robustness, which is easy to combine with other methods. Excellent distributed computing mechanism and other advantages. In this paper, the enhanced honeybee algorithm is used as a search algorithm for cloud manufacturing resource services. Finally, a simulation test environment based on QoS information cloud service discovery is built. The improved honeybee algorithm is used to test and compare its performance. The comparison experiment shows that the search algorithm in this paper is faster and more stable in the convergence time and accuracy.
【学位授予单位】：武汉理工大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.09
【参考文献】
相关期刊论文 前10条
1 孙其博;刘杰;黎
本文编号：2406481
本文链接：https://www.wllwen.com/guanlilunwen/ydhl/2406481.html
