东篱科研大数据发现系统（DRDS）

位置：成果数据库 > 期刊 > 期刊详情页

TCLM—P2P：面向P2P社区的任务协作逻辑模型

ISSN号：1000-1239
期刊名称：《计算机研究与发展》
时间：0
分类：TP393.02[自动化与计算机技术—计算机应用技术;自动化与计算机技术—计算机科学与技术]
作者机构：[1]安徽师范大学数学计算机学院,安徽芜湖241000, [2]南京邮电大学计算机学院,南京210003, [3]中国科学技术大学计算机科学与技术学院,合肥230027
相关基金：国家自然科学基金项目（60973139,60903181,61003039,61003236）;中国博士后基金项目（20100471356,20100480701）;教育部人文社科青年基金项目（11YJC880119）

作者：王杨[1,2], 王汝传[1], 严远亭[1], 韩志杰[2], 赵保华[3]

关键词：对等网络, 智能代理, 虚拟积分, 激励, 合同网, P2P, agent, virtual score, incentive, contract net

中文摘要：

P2P网络中广泛存在的“free riding”现象使其在任务协作领域的应用受到了极大制约．为了实现P2P网络环境下的有效任务协作，提出了一种具有激励机制的任务协作逻辑模型．基于Agent理论，首先给出了对等体、半对等体、P2P社区等概念；然后在合同网的框架下提出了面向P2P网络社区的任务协作逻辑模型TCLM—P2P（task collaborative logic model oriented to P2P community）．相对于传统的任务协作模型，在合理的前提假设条件下，模型给出了模型公理和协作规则．该模型通过基于虚拟积分的协作算法实现了具有激励机制的P2P网络中的任务分配与协作．原型系统的实现及仿真实验结果表明TCLM—P2P模型具有可行性和有效性：不仅能够激励自利节点主动参与到任务分配与协作中；同时也能在一定程度上抑制节点的free riding行为，从而保障了P2P系统的有序工作．

英文摘要：

Traditional P2P networks mainly are applied to file sharing and instant message fields. However, how to perform the task collaboration based on P2P community is a challenging job. The former research work indicated that the task collaboration in P2P network had been greatly restricted by free riding behaviors. To realize effective task allocating and task collaborating in P2P network environment, this paper presents a task collaboration logic model oriented to P2P community. Based on agent and multi-agent theory, the paper firstly introduces some concepts including the peer body, half-peer body and P2P community; then the TCLM-P2P is presented including some collaboration axioms and rulers. In order to enhance the incentive mechanism, virtual score becomes the main goal which each peer endeavor pursues. In addition, based on the contract net protocol, a task collaboration algorithm is presented. The proposed algorithm is composed of two phases. One is the task collaboration and the other is the task second bid when some peers fail to complete the former task. Compared with the traditional task collaboration models, the presented model has the feasible incentive mechanism in the process of task allocation and collaboration. The developed prototype and simulation results indicate that TCLM-P2P model is feasible and effective. It could not only incentivize self-interested peer to participate in task allocation and collaboration, but also restrain peer＇s free riding behaviors in some degree.

同期刊论文项目