【作者】 皇甫彩虹；

【导师】 沈记全；

【作者基本信息】 河南理工大学 ， 计算机应用技术， 2009， 硕士

【摘要】 随着Web服务近年来的蓬勃发展,越来越多稳定易用的Web服务在网络上共享。单个Web服务已经不能满足实际需求,需要将网络上的Web服务组合起来,提供更为强大的服务功能。如何快速、有效地从众多服务中找出最优的服务,己经成为现在迫切需要得到解决的问题。本文在详细分析、研究当前较为流行的BPEL4WS、eFlow、SWORD、SELF-SERV等Web服务组合方法的基础上,发现它们或多或少存在算法复杂、效率不高、并行度低等缺点,进而提出了一种效率和并行度较高的基于蚁群算法的服务组合方法。论文首先论述了Web服务组合及其关键技术,详细介绍了蚁群算法的基本原理、基本模型与特点。然后,通过对服务组合问题的研究,发现服务组合问题可以转化为联盟生成问题来解决,而蚁群算法在解决联盟生成问题上已经有了广泛的应用,因此本文应用蚁群算法来解决Web服务组合问题。文章从算法模型建立、模型参数定义、信息素的初始值确定、信息素更新策略改进、服务匹配机制建立、算法流程设计等方面来设计和实现基于蚁群算法的服务组合方法。最后,对算法的性能给出了评价,蚁群算法有效地解决了服务组合中算法复杂,效率低等缺点。更多还原

【Abstract】 With Web service developed in recent years, more and more stable-to-use Web serivices share on the network. But, individual Web service can no longer meets the actual demand, we need together web services in the network to provide a more powerful service functions. However, how quickly and effectively from many services to find the most suitable services, which has been a urgent question need to be resolved. On the basis of detailed analysising and researching Web service combination of methods such as BPEL4WS, eFlow, SWORD, SELF-SERV, this paper finds these methods more and less exist complicated algorithm, inefficient and low degree of parallelism, so brings up a kind of high efficient and high degree of parallelism service composition method based on ant colony algorithm.First, this paper discusses Web services composition and its key technologies, and detailed introduces the basic principles of the ant colony algorithm, the basic model and the characteristics. Then, by researching of Web service composition method, we find the problem of service composition can be solved by translating into the coalition generation problem. The paper from seting up algorithm model, defining the parameters of the model, determining the initial value of the pheromone, improving the renewal strategy, seting up services matching mechanism, and designing algorithm flow to design and achieve the service composition method based on ant colony algorthm. Finally, the algorithm gives the performance evaluation: the ant colony algorithm effectively solves the complex algorithm, low efficiency of the shortcomings in the services composition.更多还原