电子学报 ›› 2015, Vol. 43 ›› Issue (11): 2145-2150.DOI: 10.3969/j.issn.0372-2112.2015.11.002

• 学术论文 • 上一篇    下一篇

一种基于QoE量化评估的服务选择算法

李春林1, 廖丹2, 熊玲3, 黄月江1   

  1. 1. 中国电子科技集团公司第三十研究所, 四川 成都 610041;
    2. 西南交通大学, 四川 成都 610031;
    3. 电子科技大学, 四川 成都 611731
  • 收稿日期:2014-05-08 修回日期:2014-12-17 出版日期:2015-11-25 发布日期:2015-11-25
  • 通讯作者: 李春林
  • 作者简介:廖丹 男,1980年7月出生于重庆市,现为电子科技大学副教授,工学博士,研究方向为下一代互联网,网络优化设计.E-mail:liaodan@uestc.edu.cn
  • 基金资助:

    国家"973"重点基础研究发展计划(No.2013CB329103);国家自然科学基金(No.61201129,No.61271171,No.61001084);长江学者和创新团队发展计划及高等学校学科创新引智计划资助(No.B14039)

A Service Selection Algorithm Based on Quantified QoE Evaluation

LI Chun-lin1, LIAO Dan2, XIONG Ling3, HUANG Yue-jiang1   

  1. 1. No.30 Institute, China Electronics Technology Group Corporation, Chengdu, Sichuan 610041, China;
    2. Southwest Jiaotong University of China, Chengdu, Sichuan 610031, China;
    3. University of Electronic Science and Technology of China, Chengdu, Sichuan 611731, China
  • Received:2014-05-08 Revised:2014-12-17 Online:2015-11-25 Published:2015-11-25

摘要:

针对如何为互联网用户从多个相同或相似的服务中进行选择的问题,提出了一种新的服务选择算法:基于QoE(Quality of Experience)量化评估的服务选择算法(A Service Selecting Algorithm Based on Quantified QoE Evaluation,ASSABQ).该算法基于一种层次化评分模型,从历史评分中学习获取用户偏好,根据多种评价因素计算每个可用服务的满意度,并选择满意度最高的服务给用户.与已知算法相比,ASSABQ算法的复杂度从O(n2)下降到O(n).仿真实验结果表明,在相同应用场景下,采用ASSABQ算法得到的用户满意度比已知算法提高约10%.

关键词: 服务选择, 层次化模型, 体验质量, 用户偏好

Abstract:

Aiming at the problem of selecting a better service with the same or similar functions,one method is considering the users' experiences on the service quality.Then an algorithm called A Service Selecting Algorithm Based on Quantified QoE Evaluation(ASSABQ) is proposed.The algorithm is deduced from a hierarchy rating model and the corresponding user preferences.User preferences can be acquired from analyzing the statistical data of users' satisfaction feedbacks.Each service is rated with a satisfaction score according to the algorithm.The highest scored service will be selected to the user.Experiments compare the algorithm with both different configurations and algorithms.Compared to FCE_RS algorithm,the algorithm complexity of ASSABQ is improved from O(n2) to O(n),and the user satisfaction ratio is elevated by 10% in the same simulation scenario.

Key words: service selection, hierarchy model, quality of experience(QoE), user preference

中图分类号: