电子学报 ›› 2017, Vol. 45 ›› Issue (5): 1150-1157.DOI: 10.3969/j.issn.0372-2112.2017.05.018

• 学术论文 • 上一篇    下一篇

基于时序约束分解的QoS感知的Web服务组合

叶恒舟1,2, 李陶深3, 关云慧2   

  1. 1. 广西大学电气工程学院, 广西南宁 530004;
    2. 桂林理工大学信息科学与工程学院, 广西桂林 541000;
    3. 广西大学信息与工程学院, 广西南宁 530004
  • 收稿日期:2015-10-20 修回日期:2016-05-06 出版日期:2017-05-25
    • 作者简介:
    • 叶恒舟 男,1980年出生于湖北汉川,广西大学电气工程学院博士研究生,桂林理工大学信息科学与工程学院副教授,研究方向:Web服务组合、优化设计.E-mail:18123408@qq.com;李陶深 男,1957年出生于广西南宁,广西大学计算机与电子信息学院教授,博士生导师,研究方向:CAD理论与应用、分布式工程数据库、遗传优化设计、网络计算与信息安全、无线Mesh网络.E-mail:tshli@gxu.edu.cn;关云慧 女,1990年出生于安徽安庆,桂林理工大学硕士研究生,研究方向:Web服务技术、组合优化.
    • 基金资助:
    • 国家自然科学基金 (No.51365010); 广西自然科学基金 (No.2014GXNSFBA118269)

QoS-Aware Web Service Composition Based on Temporal Constraints Decomposition

YE Heng-zhou1,2, LI Tao-shen3, GUAN Yun-hui2   

  1. 1. College of Electrical Engineering, Guangxi University, Nanning, Guangxi 530004, China;
    2. College of Information Science and Engineering, Guilin University of Technology, Guilin, Guangxi 541000, China;
    3. College of Information and Engineering, Guangxi University, Nanning, Guangxi 530004, China
  • Received:2015-10-20 Revised:2016-05-06 Online:2017-05-25 Published:2017-05-25

摘要:

基于时序约束的QoS感知的Web服务组合(TC_QSC)问题是在考虑时序约束的基础上寻找满足QoS约束或效用最大化的Web服务组合问题,受到了越来越多的关注.本文提出了一种时序约束分解方法,把施加于整个或部分工作流的时序约束分解为施加于每个活动的局部时序约束,从而将TC_QSC问题转换为一般的QoS感知的Web服务组合(QSC)问题,并通过过滤不满足局部时序约束的候选服务,一定程度上减小原问题的规模.这种时序约束分解过程主要依赖于工作流及其涉及的活动,而与各活动的候选服务关联不大,复杂度较低.实验测试了该方法的效果与时间开销,验证了其对于局部优选算法的必要性.

关键词: 时序约束, 约束分解, QoS感知, Web服务组合, 贪心算法

Abstract:

The problem of Temporal Constraints based QoS-aware web Service Composition (TC_QSC) is becoming increasingly critical in recent years.Its objective is to select the Web service composition which maximizes the QoS utility while preserving QoS constraints and temporal constraints.This paper proposes a temporal constraints decomposition method that formulates the TC_QSC problem into a general QoS-aware Web service composition (QSC) problem by means of decomposing the temporal constraints imposed on the whole or part of the workflow to a set of local temporal constraints for each activity.So the problem scale is reduced by filtering out those candidate web services violating local temporal constraints.The temporal constraints decomposition process is mainly dependent on the workflow and activities involved in it,while it has little to do with the candidate services.Therefore,its complexity is lower.The experiments have demonstrated the effectiveness and time cost of the proposed method and verify its necessity to local selection algorithm.

Key words: temporal constraint, constraint decomposition, QoS-aware, Web service composition, greedy algorithm

中图分类号: