基于模糊Petri网的服务组合优化技术研究与实现

发布时间：2018-04-28 08:08

本文选题：服务组合优化 + QoS　；参考：《沈阳理工大学》2017年硕士论文

【摘要】：互联网技术的不断发展,使人们越来越重视对Web服务的研究以及应用。然而,现在Web服务因具有松散耦合、高度可集成、跨平台性等特性不断发展的同时,拥有相同或者相似功能的服务数量却不断增多。这些Web服务由于粒度较小、功能简单无法满足用户复杂需求,所以对Web服务进行组合优化显得尤为重要。本文首先对当前服务描述模型的优缺点进行研究,提出QoS描述模型。单个QoS模型包括服务成本、执行时间、可用性、安全性、信誉度五个属性。在对单个QoS属性进行量化的基础上,引入权重的概念,对QoS进行综合评价。服务组合优化的QoS计算模型是在单个QoS综合评价的基础上,根据服务组合优化的流程结构,对服务的QoS属性进行计算。其次,为了提高服务组合优化的效率和准确性,对标准粒子群算法的惯性权重和学习因子进行改进,提出一种线性和非线性结合的惯性权重递减策略,并且结合改进的三角函数变化的学习因子,有效的提高了粒子群算法的收敛性。将改进的粒子群算法应用到服务组合优化中,提高了服务组合优化的效率。最后,研究模糊Petri网的相关知识,把基于改进粒子群的服务组合优化算法与模糊Petri网相结合。构建服务实例,生成服务依赖关系生成图,用库所表示一类的抽象服务,每个库所对应一定数量的具体服务,对这些服务进行服务组合优化,选择出一组满足需求的服务,并且开发一个基于模糊Petri网的服务组合优化的系统。
[Abstract]:With the development of Internet technology, people pay more and more attention to the research and application of Web services. However, with the development of loosely coupled, highly integrated and cross-platform Web services, the number of services with the same or similar functions is increasing. Because of their small granularity and simple functions, these Web services can not meet the complex needs of users, so it is very important to optimize the composition of Web services. In this paper, the advantages and disadvantages of the current service description model are studied, and the QoS description model is proposed. A single QoS model includes five attributes: service cost, execution time, availability, security, and reputation. Based on the quantization of single QoS attribute, the concept of weight is introduced to evaluate QoS synthetically. The QoS computing model of service composition optimization is based on the comprehensive evaluation of a single QoS, according to the process structure of service composition optimization, the QoS attribute of the service is calculated. Secondly, in order to improve the efficiency and accuracy of service composition optimization, the inertia weight and learning factor of standard particle swarm optimization are improved, and a linear and nonlinear inertia weight decreasing strategy is proposed. Combined with the improved trigonometric function change learning factor, the convergence of PSO is improved effectively. The improved particle swarm optimization algorithm is applied to service composition optimization to improve the efficiency of service composition optimization. Finally, the related knowledge of fuzzy Petri nets is studied, and the service composition optimization algorithm based on improved particle swarm optimization is combined with fuzzy Petri nets. The service instance is constructed, the service dependency graph is generated, the abstract services are represented by the library, each library corresponds to a certain number of specific services, and optimizes the service composition of these services, and selects a set of services to meet the requirements. A service composition optimization system based on fuzzy Petri net is developed.
【学位授予单位】：沈阳理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP393.09;TP301.1

【参考文献】