基于改进布谷鸟算法的Web服务组合QoS最优化问题研究
发布时间：2018-10-26 20:13
【摘要】：随着近年来云计算技术的蓬勃发展，使得Web服务的研究和应用迎来了新的挑战。而由于Web服务数量的增加，Web服务组合问题的研究也得到了广泛的关注。传统的Web服务组合一般只考虑服务的功能性需求，对服务的非功能性需求和总体服务质量（QoS）考虑较少。可以说，服务选择算法的选取是关系到最终决定服务组合QoS性能好坏的关键性因素。 首先本文对近年来基于QoS的服务组合问题从多个研究方向进行了分类和总结。在选择优化算法的选择上，普遍采用粒子群算法、蚁群算法、Skyline方法等。因大量实验已经证明布谷鸟算法在解决多目标函数方面的表现要优于其他现存的算法，而且相较于其他的元启发式优化算法具有更好的通用性,因此作者提出将其应用于服务组合问题。本文首先建立了将QoS的服务组合问题从多目标问题转化为单目标问题的数学模型。通过使用从QWS数据库中选定的相似服务，，构建抽象服务组合模型。为了验证该方法的可行性，本文还在提出的数学模型的基础上实现了目前研究最广泛的基于粒子群算法的Web服务组合最优化问题，通过对比，最终证明了布谷鸟算法在解决Web服务组合问题上的可行性。最后对原始布谷鸟算法的随机取值过程进行了改进，并进一步通过实验验证，改进的布谷鸟算法在Web服务组合最优化问题的实现上效率的提高。
[Abstract]:With the rapid development of cloud computing technology in recent years, the research and application of Web services is facing new challenges. Due to the increasing number of Web services, the research on Web services composition has been paid more and more attention. The traditional Web service composition only considers the functional requirements of the service, but the non-functional requirements and the overall quality of service (QoS) are less considered. It can be said that the selection of service selection algorithm is the key factor to determine the performance of service composition QoS. Firstly, this paper classifies and summarizes the service composition problems based on QoS in recent years. Particle swarm optimization (PSO), ant colony algorithm (ACO) and Skyline method are widely used in the selection of optimization algorithm. Because a large number of experiments have proved that the cuckoo algorithm is superior to other existing algorithms in solving multi-objective functions, and compared with other meta-heuristic optimization algorithms, it has better universality. Therefore, the author proposes to apply it to the service composition problem. In this paper, a mathematical model is established to transform the service composition problem of QoS from multi-objective problem to single-objective problem. An abstract service composition model is constructed by using similar services selected from the QWS database. In order to verify the feasibility of the proposed method, this paper also implements the most widely studied Web service composition optimization problem based on particle swarm optimization algorithm on the basis of the proposed mathematical model. Finally, the feasibility of cuckoo algorithm in solving Web service composition problem is proved. Finally, the random selection process of the original cuckoo algorithm is improved, and further verified by experiments, the efficiency of the improved cuckoo algorithm in the Web service composition optimization problem is improved.
【学位授予单位】：东北师范大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.09;TP301.6
本文编号：2296827
[Abstract]:With the rapid development of cloud computing technology in recent years, the research and application of Web services is facing new challenges. Due to the increasing number of Web services, the research on Web services composition has been paid more and more attention. The traditional Web service composition only considers the functional requirements of the service, but the non-functional requirements and the overall quality of service (QoS) are less considered. It can be said that the selection of service selection algorithm is the key factor to determine the performance of service composition QoS. Firstly, this paper classifies and summarizes the service composition problems based on QoS in recent years. Particle swarm optimization (PSO), ant colony algorithm (ACO) and Skyline method are widely used in the selection of optimization algorithm. Because a large number of experiments have proved that the cuckoo algorithm is superior to other existing algorithms in solving multi-objective functions, and compared with other meta-heuristic optimization algorithms, it has better universality. Therefore, the author proposes to apply it to the service composition problem. In this paper, a mathematical model is established to transform the service composition problem of QoS from multi-objective problem to single-objective problem. An abstract service composition model is constructed by using similar services selected from the QWS database. In order to verify the feasibility of the proposed method, this paper also implements the most widely studied Web service composition optimization problem based on particle swarm optimization algorithm on the basis of the proposed mathematical model. Finally, the feasibility of cuckoo algorithm in solving Web service composition problem is proved. Finally, the random selection process of the original cuckoo algorithm is improved, and further verified by experiments, the efficiency of the improved cuckoo algorithm in the Web service composition optimization problem is improved.
【学位授予单位】：东北师范大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.09;TP301.6
【参考文献】
相关期刊论文 前8条
1 苏森;李飞;杨放春;;分布式环境中服务组合的迭代选择算法[J];中国科学(E辑:信息科学);2008年10期
2 张佩云;孙亚民;;动态Web服务组合研究[J];计算机科学;2007年05期
3 夏亚梅;程渤;陈俊亮;孟祥武;刘栋;;基于改进蚁群算法的服务组合优化[J];计算机学报;2012年02期
4 刘长平;叶春明;;求解置换流水车间调度问题的布谷鸟算法[J];上海理工大学学报;2013年01期
5 冯登科;阮奇;杜利敏;;二进制布谷鸟搜索算法[J];计算机应用;2013年06期
6 董元元;倪宏;邓浩江;刘磊;;QoS全局最优化的服务选择策略[J];小型微型计算机系统;2011年03期
7 欧伟杰;曾承;曾青;彭智勇;王珍珍;刘波;马景燕;;QoS感知的高效抽象服务选择[J];小型微型计算机系统;2013年01期
8 冯建周;孔令富;;基于QoS的Web服务组合中多路径全局优化方法的研究[J];小型微型计算机系统;2013年07期
本文编号：2296827
本文链接：https://www.wllwen.com/guanlilunwen/ydhl/2296827.html
