基于非线性规划的Web服务组合优化研究

发布时间：2018-04-04 17:53

  本文选题：非线性规划　切入点：Web服务组合　出处：《江西师范大学》2014年硕士论文

【摘要】：随着Web服务的快速发展，，基于Qos的Web服务组合优化问题已经成为Web服务的研究热点。Web服务组合是NP-hard问题。典型的与Web服务相关的Qos的属性有响应时间、吞吐率、到达率等。对基于Qos的Web服务组合优化问题的研究，可以更全面的满足用户对服务质量的要求，这对于Web服务的发展具有十分重要的意义。 本文首先介绍了目前求解Web服务组合优化问题常用的几个算法。主要有遗传算法、粒子群算法等等。遗传算法在求解Web服务优化问题中应用广泛，但其参数选取大多依靠经验，不能保证解的有效性，并且在最优解的搜索过程中容易陷入局部收敛。蚁群算法全局搜索能力强，但其模型并不能直接应用于Web服务组合的实际优化问题。 本文将BPEL服务组合流程映射为排队网络，利用排队网络的理论对响应时间、吞吐率以及到达率等Qos属性进行分析，并采用Mathematica对数据进行计算。计算结果表明，该方法具有很强的实用性。 本文的研究内容是在给定的Web服务中利用非线性规划算法优选出一组Web服务，使得这组服务在满足约束条件的前提下总费用达到最小。本文基于排队网络理论，将BPEL顺序结构、BPEL条件结构以及BPEL循环结构映射为排队网络，利用排队网络的特点对三种结构下Web服务的吞吐率进行计算，进而求解Web服务器的响应时间关于节点i (1≤i≤N)的服务率的表达式。通过对响应时间以及各个节点的服务率设置约束条件，建立非线性规划模型，在Mathematica中对节点i (1≤i≤N)的服务率进行求解。通过将求解的服务率与给定的Web服务的服务率相比较，选取与所求解的服务率相差最小的一组Web服务，就是本文所求的最优Web服务组合。
[Abstract]:With the rapid development of Web services, the optimization of Web services composition based on Qos has become the research hotspot of Web services.Typical Qos attributes associated with Web services are response time, throughput, arrival rate, and so on.The research on the optimization of Web services composition based on Qos can more fully meet the requirements of users for the quality of service, which is of great significance for the development of Web services.This paper first introduces several algorithms commonly used to solve Web service composition optimization problem.There are genetic algorithm, particle swarm optimization algorithm and so on.Genetic algorithm (GA) is widely used in solving Web service optimization problem, but its parameter selection depends on experience, which can not guarantee the validity of the solution, and it is easy to fall into local convergence in the process of searching for the optimal solution.Ant colony algorithm has strong global search ability, but its model can not be directly applied to the practical optimization problem of Web service composition.In this paper, the BPEL service composition flow is mapped to a queueing network, and the response time, throughput and arrival rate are analyzed by using the queuing network theory, and the data are calculated by Mathematica.The calculation results show that the method is very practical.The purpose of this paper is to select a set of Web services by using nonlinear programming algorithm in a given Web service, so that the total cost of the set of services can be minimized under the premise of satisfying the constraint conditions.Based on the queuing network theory, the BPEL sequential structure and BPEL cyclic structure are mapped to queueing network, and the throughput of Web services under three kinds of structures is calculated by using the characteristics of queueing network.Then the expression of the response time of the Web server to the service rate of node I 1 鈮

本文编号：1711037