奇异方向上两人零和博弈的近似纳什均衡求解
Approximate Nash Equilibrium Solution of Two Person Zero Sum Game in Singular Direction
DOI: 10.12677/AAM.2022.1111866, PDF,    科研立项经费支持
作者: 龚谊承*:武汉科技大学理学院数学与统计系,湖北 武汉;科技大学冶金工业过程系统科学湖北省重点实验室,湖北 武汉;吴雯珑:武汉科技大学理学院数学与统计系,湖北 武汉
关键词: 非方支付矩阵两人零和博弈纳什均衡解奇异向量Non-Square Payoff Matrix Two-Person Zero-Sum Game Nash Equilibrium Solution Singular Vector
摘要: 纳什均衡的近似求解是博弈论中基础且重要的话题,其中基于特征向量的方法提供了一个新的视角,因此对博弈矩阵非方阵情形的探讨是值得期待的。本文以两人零和博弈为例,探讨了纳什均衡求解体系与非方支付矩阵的奇异方向的关系。关于纳什均衡计算的两个定理表明:当支付矩阵存在非负奇异值,且其对应的左奇异向量和右奇异向量元素都非负时,则该奇异向量分别对应博弈双方的纳什均衡解。以局中人的策略选择数目分别为3个和2个的情况为算例,验证了该定理的适用性,并对大规模非方矩阵博弈的纳什均衡的近似求解提供了一个新的方向。
Abstract: The approximate solution of Nash equilibrium is a basic and important topic in game theory, in which the method based on the eigenvector provides a new perspective, so it is worth looking forward to the discussion of the non-square matrix of the game. Taking a two-person zero-sum game as an example, this paper discusses the relationship between Nash equilibrium solution system and the singular direction of the non-square payment matrix. Two theorems about Nash equilibrium calculation show that when the payoff matrix has a non-negative singular value, and its corresponding left singular vector and right singular vector elements are all non-negative, then the singular vectors correspond to the Nash equilibrium solutions of both players in the game. Taking the cases where the number of strategy choices of the players in the game is 3 and 2 respectively, the applicability of the theorem is verified, and a new direction is provided for the approximate solution of Nash equilibrium for large-scale non-square matrix games.
文章引用:龚谊承, 吴雯珑. 奇异方向上两人零和博弈的近似纳什均衡求解[J]. 应用数学进展, 2022, 11(11): 8183-8190. https://doi.org/10.12677/AAM.2022.1111866

参考文献

[1] 谢炽予. 经济博弈论[M]. 第4版. 上海: 复旦大学出版社, 2018: 32-35.
[2] 安京京, 南江霞, 卜红. 模糊二人零和对策的纳什均衡求解[J]. 经济数学, 2015, 32(3): 36-40.
[3] 刘豪. 多智能体博弈强化学习算法及其均衡研究[D]: [硕士学位论文]. 陕西: 西安科技大学, 2020: 1-3.
[4] Sadana, U., Reddy, P.V. and Georges, Z. (2021) Nash Equilibria in Nonzero-Sum Differential Games with Impulse Control. European Journal of Operational Research, 295, 792-805. [Google Scholar] [CrossRef
[5] Ganzfried, S. (2021) Algorithm for Computing Ap-proximate Nash Equilibrium in Continuous Games with Application to Continuous Blotto. Games, 12, Article No. 47. [Google Scholar] [CrossRef
[6] 侯剑, 李萌萌, 文竹. 一类纳什均衡问题的求解算法[J/OL]. 运筹学学报: 1-8. http://kns.cnki.net/kcms/detail/31.1732.O1.20220505.1043.018.html, 2022-10-09.
[7] 袁唯淋, 罗俊仁, 陆丽娜, 等. 智能博弈对抗方法:博弈论与强化学习综合视角对比分析[J]. 计算机科学, 2022, 49(8): 191-204.
[8] Liu, L. and Jia, W. (2021) The Value Function with Regret Minimization Algorithm for Solving the Nash Equilibrium of Mul-ti-Agent Stochastic Game. International Journal of Computational Intelligence Systems, 14, 1633-1641. [Google Scholar] [CrossRef
[9] Weil Jr., R.L. (1968) Game Theory and Eigensystems. SIAM Re-view, 10, 360-367. [Google Scholar] [CrossRef
[10] Thompson, G.L. and Weil Jr., R.L. (1969) Further Rela-tions between Game Theory and Eigensystems. SIAM Review, 11, 597-602. [Google Scholar] [CrossRef
[11] Wernick, R.J. (1970) Parametrized Games and the Eigenproblem. Linear Algebra and Its Applications, 3, 311-346. [Google Scholar] [CrossRef
[12] Liu, H. (2020) The Relation between Eigenvalue/Eigenvector and Matrix Game. ArXiv: 2012.01269.
[13] 王照琨. 二阶与三阶收益矩阵博弈的纳什均衡的算法[D]: [硕士学位论文]. 四川: 四川师范大学, 2019: 2-4.
[14] Gemp, I., McWilliams, B., Vernade, C. and Graepel, T. (2020) Eigengame: PCA as a Nash Equilibrium. ArXiv: 2010.00554.
[15] Szabó, G. and Király, B. (2021) General Features of Nash Equilibria in Combinations of Elementary Interactions in Symmetric Two-Person Games. The European Physical Journal B, 94, Article No. 102. [Google Scholar] [CrossRef
[16] 刘昊天, 王玉惠, 陈谋, 张逸航. 基于对局迭代的无人机空战博弈研究[J]. 电光与控制, 2022, 29(2): 1-6.
[17] Fudenberg, D. and Tirole, J. (1991) Game Theory. The MIT Press, Cambridge, 80-98.
[18] Nash Jr., J.F. (1950) Equilibrium Points in n-Person Games. Proceedings of the National Academy of Sciences of the United States of America, 36, 48-49. [Google Scholar] [CrossRef] [PubMed]
[19] 于维生. 博弈论与经济[M]. 北京: 高等教育出版社, 2007: 31-50.
[20] Thompson, G.L. and Weil, R.L. (1971) Von Neumann Model Solutions Are Generalized Eigensystems. In: Bruckmann, G., Ed., Contributions to the Von Neumann Growth Model, Springer, Berlin, 139-154. [Google Scholar] [CrossRef
[21] Giorgi, G. (2016) Eigenvalues and Eigenvectors in Von Neumann and Related Growth Models: An Overview and Some Remarks. Journal of Mathematics Research, 8, 24-37. [Google Scholar] [CrossRef