arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2406.03565cs.GTcs.MAcs.SYeess.SY
收藏

在零和博弈中寻找局部纳什均衡的二阶算法

Second-Order Algorithms for Finding Local Nash Equilibria in Zero-Sum Games

The University of Texas at Austin(德克萨斯大学奥斯汀分校);MIT Lincoln Laboratory(麻省理工学院林肯实验室)

Kushagra Gupta, Xinjie Liu, Ross Allen, Ufuk Topcu, David Fridovich-Keil

AI总结:

针对零和博弈中算法易收敛至非纳什驻点的问题,提出基于动力系统理论的二阶算法,保证仅收敛到局部纳什均衡并给出线性或超线性收敛率,且可推广至耦合约束情形。

摘要:

零和博弈出现在各种各样的问题中,包括鲁棒优化和对抗学习。然而,用于在这些博弈中寻找局部纳什均衡的算法常常会收敛到非纳什驻点。这凸显了一个关键挑战:对于任何算法,其底层动力系统的稳定性属性都可能使非纳什点成为潜在吸引子。为克服这一挑战,算法必须考虑与参与者代价曲率有关的细微之处。为此,我们利用动力系统理论,开发了一种二阶算法,用于在光滑、可能非凸-非凹的零和博弈设定中寻找局部纳什均衡。首先,我们证明这种新方法保证仅收敛到局部纳什均衡,并具有渐近局部线性收敛率。随后,我们将该方法的一个版本解释为一种改进的 Gauss-Newton 算法,它在满足一阶局部纳什均衡条件的点的邻域内具有局部超线性收敛性。相比之下,当前具有类似保证的相关最先进方法在非凸-非凹设定中并不提供收敛率。此外,我们表明该方法可自然推广到具有凸约束且可能存在耦合约束的设定,同时保留此前仅收敛到局部(广义)纳什均衡的保证。我们实验的代码可在 https://github.com/CLeARoboticsLab/ZeroSumGameSolve.jl 找到。

英文摘要:

Zero-sum games arise in a wide variety of problems, including robust optimization and adversarial learning. However, algorithms deployed for finding a local Nash equilibrium in these games often converge to non-Nash stationary points. This highlights a key challenge: for any algorithm, the stability properties of its underlying dynamical system can cause non-Nash points to be potential attractors. To overcome this challenge, algorithms must account for subtleties involving the curvatures of players' costs. To this end, we leverage dynamical system theory and develop a second-order algorithm for finding a local Nash equilibrium in the smooth, possibly nonconvex-nonconcave, zero-sum game setting. First, we prove that this novel method guarantees convergence to only local Nash equilibria with an asymptotic local \textit{linear} convergence rate. We then interpret a version of this method as a modified Gauss-Newton algorithm with local \textit{superlinear} convergence to the neighborhood of a point that satisfies first-order local Nash equilibrium conditions. In comparison, current related state-of-the-art methods with similar guarantees do not offer convergence rates in the nonconvex-nonconcave setting. Furthermore, we show that this approach naturally generalizes to settings with convex and potentially coupled constraints while retaining earlier guarantees of convergence to only local (generalized) Nash equilibria. Code for our experiments can be found at https://github.com/CLeARoboticsLab/ZeroSumGameSolve.jl.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑