Web服务中若干问题的研究

发布时间：2018-04-19 07:13

本文选题：Web服务 + Web服务选取　；参考：《吉林大学》2014年博士论文

【摘要】：本文主要研究内容为当前Web服务领域中的几个热点问题，包括Web服务选取方法、多等级Web服务部署问题以及Web服务组合运行时容错处理和异常恢复方法及策略的研究。本文中首先提出了一种改进的人工蜂群算法用来求解服务选取问题。结合遗传操作设计雇用蜂觅食及侦查蜂策略，并针对服务选取问题的特点提出了基于效用值及违反约束度的食物源评价方法及基于邻域的随机贪心观察蜂觅食策略，进而建立求解该问题的人工蜂群算法优化模型，在此基础上实现了求解服务选取问题的混合人工蜂群算法-GABC。其次提出了一种基于协同过滤推荐的混合式Web服务选取方法，将Web服务请求者的个人兴趣偏好和相似用户群体的经验相结合，利用推荐技术对具有相同功能和不同Qos属性值的Web服务进行选取。再次提出了一种改进的遗传算法用来求解多等级服务部署问题。在选择过程中引入了个体被支配强度，通过将个体的支配强度和被支配强度结合到一起建立对个体的评价策略，进而对效用函数进行重新定义，，并设计了基于概率的交叉策略及结合局部搜索的个体变异策略和根据评价结果进行环境选择及生成个体的交叉概率方法。此外还设计新的局部搜索策略并将其融入到变异策略中，提高了变异操作的有效性。最后给出了一种基于服务冗余和约束冗余的Web服务组合运行时容错处理和异常恢复策略。通过在组合规划阶段为服务流程中每个抽象成员服务建立后备服务组和将约束分解到每个抽象成员服务上的方法，为运行时的容错处理和异常恢复提供了一种以尽可能小的代价恢复服务组合运行并尽量保留已执行部分结果的处理策略。在此基础上，进一步给出了服务组件库的建立和维护策略及其算法。本文中针对Web服务领域中几个关键问题进行了研究和探讨，所做的主要研究工作和研究成果在提高web服务和服务组合的可靠性和效率等方面具有一定的意义。
[Abstract]:This paper focuses on several hot topics in the field of Web services, including the selection of Web services, the deployment of multilevel Web services, and the research of fault tolerance and exception recovery methods and strategies in Web service composition runtime.
This paper first presents an improved artificial bee colony algorithm is used to solve the problem of the service selection. Combined with the genetic operators employed bee and bee foraging investigation strategy, according to the characteristics of the service selection problem and put forward the utility value and evaluation method of constraint violation of food source and neighborhood random greedy bee foraging strategy based on observation, optimization model and the establishment of artificial bee colony algorithm for solving this problem, based on the realization of the -GABC. hybrid artificial bee colony algorithm selection problem solving service
Secondly, a hybrid Web service selection method based on collaborative filtering recommendation is proposed, which combines the Web service requester's personal interest preference with the experience of similar user groups, and uses recommendation technology to select Web services with the same function and different Qos attribute values.
Again, this paper proposes an improved genetic algorithm for solving multi class service deployment problem. In the selection process using individual dominated by strength, will dominate and strength of individual dominated strength combined to establish evaluation strategies for individuals together, and the utility function for re defined and designed based on the crossover probability and combined with the local search strategy according to the individual variation and crossover probability evaluation results of environmental selection and generation of the individual. In addition, the design of a new local search strategy and put it into the mutation strategy, the effectiveness of high mutation operation.
Finally, a Web service composition service redundancy and redundant constraints based on the fault-tolerant processing and exception recovery strategy. Through each service process Abstract member services build up a reserve service group and the Constraint Decomposition to each service for abstract members in combinatorial planning stage, to run the fault-tolerant processing and exception recovery a recovery service combination operation and to retain some results processing strategy has been performed with as little as possible cost. On this basis, further gives the service component database establishment and the maintenance strategy and algorithm.
In this paper, several key issues in the field of Web services are studied and discussed. The main research work and research results are of great significance in improving the reliability and efficiency of Web services and service composition.

【学位授予单位】：吉林大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：TP393.09

【参考文献】