近日，商学院程郁琨教授与北京大学邓小铁教授共同撰写的学术论文“Tight incentive Analysis of Sybil Attacks Against the Market Equilibrium of Resource Exchange over General Networks”在博弈论领域国际顶级期刊《Games and Economic Behavior》发表。程郁琨为本研究的第一作者，江南大学商学院为第一完成单位。
《Games and Economic Behavior》是博弈论领域国际公认的顶级经济期刊，其研究领域包括但不限于博弈论、经济学、政治学、生物学、计算机科学、数学、心理学等，具有极强的学术影响力。该论文是国家自然科学基金资助项目的阶段性成果。
程郁琨教授致力于算法博弈论的研究，长期专注于网络环境下的资源共享博弈机制设计与激励分析，其成果相继发表于《中国科学基金》、运筹学领域顶级期刊《Mathematics of Operations Research》、计算机科学领域的权威期刊《IEEE Transactions on Cloud Computing》、计算经济学领域的国际顶级会议ACM Conference on Economics and Computation (EC2022)、Conference on Web and Internet Economics (WINE2017,2024)、人工智能领域的国际重要会议International Joint Conference on Artificial Intelligence (IJCAI2016)、International Conference on Autonomous Agents and Multiagent Systems (AAMAS2020, 2023)等，获得国家自然科学基金委的4项资助。
Abstract：
The BitTorrent network, a well-known Internet-scale P2P system, applies the proportional response protocol to exchange resource, where each participant contributes resource to neighbors in proportion to the amount it received in the previous round. The dynamics of this protocol is known to converge to a market equilibrium. However, an agent may manipulate this protocol by a Sybil attack to create fictitious identities and control them to gain more benefit. We apply the concept of incentive ratio, the percentage of the new utility after a Sybil attack over the benchmark, to measure the incentive of a strategic agent to play a Sybil attack, proving a tight bound of two over general networks. This finding completes the theoretical picture for incentive analysis on Sybil attacks facing the legendary tit-for-tat protocol for Internet bandwidth sharing and other applications.
文章摘要
