基于QoS的Web服务组合优化研究

发布时间：2018-03-14 04:47

本文选题：Web服务　切入点：Web服务组合　出处：《燕山大学》2014年硕士论文　论文类型：学位论文

【摘要】：随着Web服务技术以及服务组合技术的迅速发展，具有相同或者相似功能的服务数量逐渐增多，利用QoS进行服务选择已经成为Web服务组合的关键技术。现有的QoS计算模型中，QoS各个属性采用固定的权值，完全忽略了各属性实际值及其权重之间存在的相互关系，导致Web服务组合质量的降低。同时Web服务组合中的服务选择算法大多采用经典的数学优化算法，这就严重影响了Web服务组合的效率。为了解决上述问题，，本文对服务组合中关键问题进行了以下研究。首先，针对Web服务现有的QoS属性评价模型的不足，提出一种QoS变权综合评价方法，在变权综合理论的基础上，根据用户对Web服务的QoS各个属性的限制，建立变权状态向量，动态调整单个服务的QoS属性权重，提高Web服务QoS综合评价的客观性和准确性。其次，针对粒子群优化算法存在的缺陷和不足，通过调节粒子群优化算法中学习因子变化策略对收敛速度进行调整，同时动态调整惯性权重，然后根据测试函数，不断的仿真实验确定改进粒子群优化算法中的参数值，使之能够较好的适应函数的需要，提高粒子速度和位置更新的速度，从而提高Web服务组合效率。最后，根据用户对Web服务组合的要求和限制，设定相应的粒子适应度函数和QoS约束条件，然后对改进前后的粒子群优化算法和优化前后的QoS评价方法进行模拟和对比实验，与现有的QoS综合评价方法和粒子群优化算法相比较，验证改进算法的有效性。
[Abstract]:With the rapid development of Web services technology and service composition technology, the number of services with the same or similar functions is increasing. Using QoS to select services has become the key technology of Web service composition. In the existing QoS computing model, each attribute adopts a fixed weight, and the relationship between the actual values and their weights of each attribute is completely ignored. At the same time, the classical mathematical optimization algorithm is used in the service selection algorithm of Web service composition, which seriously affects the efficiency of Web service composition. In this paper, the key problems of service composition are studied as follows. First of all, in view of the deficiency of the existing QoS attribute evaluation model of Web service, this paper proposes a QoS variable weight comprehensive evaluation method. Based on the theory of variable weight synthesis, the variable weight state vector is established according to the restriction of the user to the QoS attribute of Web service. The QoS attribute weight of a single service is dynamically adjusted to improve the objectivity and accuracy of Web service QoS comprehensive evaluation. Secondly, aiming at the defects and shortcomings of particle swarm optimization algorithm, the convergence rate is adjusted by adjusting the learning factor change strategy of particle swarm optimization algorithm, and the inertia weight is adjusted dynamically, and then according to the test function, Continuous simulation experiments determine the parameters of the improved particle swarm optimization algorithm, which can better meet the needs of the function, improve the speed of particle velocity and position update, and improve the efficiency of Web service composition. Finally, according to the requirements and limitations of Web service composition, the corresponding particle fitness function and QoS constraint conditions are set up, and then the particle swarm optimization algorithm and the QoS evaluation method before and after the optimization are simulated and compared. Compared with the existing QoS comprehensive evaluation method and particle swarm optimization algorithm, the effectiveness of the improved algorithm is verified.
【学位授予单位】：燕山大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.09

【参考文献】