Reputation Propagation for Web Service Composition: A Shapley Value Approach

Tao Zhang, Yongzhi Wang, Guangxia Li, Jingjing Guo, Qi LiNaNA 2017 · OpenReview

Abstract

In service-oriented computing environments, a novel challenge is to update the reputation of each composed component service. Recently, reputation propagation has become an emerging solution and received much attention in research area. Existing approaches for reputation propagation cannot capture the complex service composition structures, or cannot fairly distribute the users' subjective evaluation values. In this paper, we propose an reputation propagation approach to fairly distribute the overall subjective ratings to each components in the composition. Our approach first utilizes Beta reputation models to record the ratings of service components. Then, it computes the contribution of each component using Shapley value in corporative gaming theory, and improves the efficiency of Shapley value computation by bit vector encoding. Finally, we theoretically and experimentally validate our approach through simulated studies. The results show that our approach can hierarchically propagate the consumers' rating to each component, and can ensure fairness that no component in the composite service will receive extra rewards or punishments.
BibTeX
@inproceedings{zhang2017reputation,
  author = {Tao Zhang and Yongzhi Wang and Guangxia Li and Jingjing Guo and Qi Li},
  title = {{Reputation Propagation for Web Service Composition: A Shapley Value Approach}},
  booktitle = {NaNA},
  year = {2017},
  url = {https://openreview.net/forum?id=wWiPKTXqPU},
}

Loading related content...

Figure