zoukankan      html  css  js  c++  java
  • [论文笔记] Reputation Propagation in Composite Services (ICWS, 2009)

    Time: 2.8 hours
    Surya Nepal, Zaki Malik, Athman Bouguettaya, "Reputation Propagation in Composite Services," icws, pp.295-302, 2009 IEEE International Conference on Web Services, 2009

        作者Surya Nepal目前是CSIRO ICT centre的Principal Research Scientist, 方向是trust aspects of Service Computing. 本文要解决的问题是如何将用户针对composite web service的reputation值分配给组成服务("developing a method of distribution of reputation received by a composite service to its component services").

    1. 概念
    trust: the belief that a service consumer has about the intention and ability of a service provider to act as expected.
    reputation: a mechanism of establishing the belief about the provider's ability to deliver, through collective perception of the consumers that have interacted with the service provider in the past.

    2. Service Composition Scenario (S2)
    Vertical composition
    有点类似"centralized service composition", 有两个特点: (1) component service向composite service报告, component services之间不直接交互 (2)"sub-contracting" task of the service to other parites.
    这个第二点似乎就是指"成员服务本身也是一个composite service"的情况.
    Horizontal composition
    model a "supply-chain" like combination of services, 类似"decentralized service composition".
    Hybrid composition

    本文讨论的是vertaical composition的情形.

    3. (S3)本文提出了一种"contribution-based distribution of reputation method"来讲reputation分配给component services, 分为3步: 
    Contribution Vector Production –>  Distribution Vector Derivation –>  Reputation Propagation

    这一章节的一些疑问
    (1) 文中公式(3)缺少符号的解释. 这个公式是用来计算组成服务的reputation的, 有几个疑问: 为什么是求积? 为什么是t-1到t-k?为什么(s, C)t? where后面的那个东西是哪里冒出来的?
    image
    (2) contributor vector计算出来以后, 之后就没有用到过?
    (3) 对整套计算方法的有效性也有怀疑.

    有机会会再看一下这篇.

  • 相关阅读:
    Martin Fowler关于IOC和DI的文章(原版)
    父类引用指向子类对象详解
    求中位数总结
    二叉树的遍历方法
    MySQL知识小结
    栈和队列的基础算法学习(EPI)
    链表的基础题目学习(EPI)
    数组和字符串的基础题目学习(EPI)
    基本类型算法题目学习(EPI)
    被C语言操作符优先级坑了
  • 原文地址:https://www.cnblogs.com/yuquanlaobo/p/1608904.html
Copyright © 2011-2022 走看看