嘉宾简介：

张军欢，现任北京航空航天大学经济管理学院金融系副教授、博士生导师，金融硕士教育中心主任。英国伦敦国王学院计算机科学博士，美国佐治亚大学人工智能博士后。研究方向为人工智能和区块链在金融和经济领域的交叉应用，诸如：智能金融与区块链金融，金融科技与监管科技，算法交易与高频交易，行为与计算金融，数字资产定价与交易等。曾受邀访问美国芝加哥大学布斯商学院、范德堡大学经济系、乔治梅森大学经济系。主持国家级课题3项，包括国家重点研发计划项目课题、国家自然科学基金面上项目、青年项目等；主持北航“人文社科拔尖人才支持计划项目(A类)”“卓越百人计划”“世界顶尖大学学者合作计划”等。研究成果发表在European Journal of Finance、Quantitative Finance、Review of Quantitative Finance and Accounting、IEEE Transactions on Engineering Management、International Review of Financial Analysis等期刊。主编/联合主编《区块链金融》和《健康保险与大数据应用》。担任IMA Journal of Management Mathematics、International Review of Economics & Finance、Cogent Economics & Finance、International Journal of Financial Engineering副主编，PLOS ONE学术主编，Blockchain青年编委，IMA Journal of Management Mathematics特刊编辑；兼任中国工业与应用数学学会区块链专业委员会委员、中国智能金融产学研高峰论坛理事、中国优选法统筹法与经济数学研究会量化金融与保险分会理事。

讲座简介：

In decentralized exchanges (DEXs), sandwich attackers can collude with buildersthrough private memory pools. We study the market microstructure of automated marketmaker-based decentralized exchanges with the collusive trading strategy between sandwichattackers and builders. We explore the impact of this collusion on the behaviorsof other agents, which are arbitrageurs employing the two-point (TA) arbitrage strategy,liquidity providers using the balancing swap fees and impermanent loss strategy, swaptraders using the moving average (MA) and zero intelligence (ZI) strategies. We optimizethe trading strategies of sandwich attackers, liquidity providers, MA swap traders, andTA arbitrageurs using the multi-agent deep reinforcement learning method. Using dailyclosing prices of USDT/BNB, USDT/BTC, and USDT/ETH from January 1, 2023, toDecember 31, 2024, we construct the empirical analysis.