Papers
Topics
Authors
Recent
Search
2000 character limit reached

Learning and steering game dynamics towards desirable outcomes

Published 1 Apr 2024 in eess.SY, cs.GT, and cs.SY | (2404.01066v2)

Abstract: Game dynamics, which describe how agents' strategies evolve over time based on past interactions, can exhibit a variety of undesirable behaviours including convergence to suboptimal equilibria, cycling, and chaos. While central planners can employ incentives to mitigate such behaviors and steer game dynamics towards desirable outcomes, the effectiveness of such interventions critically relies on accurately predicting agents' responses to these incentives -- a task made particularly challenging when the underlying dynamics are unknown and observations are limited. To address this challenge, this work introduces the Side Information Assisted Regression with Model Predictive Control (SIAR-MPC) framework. We extend the recently introduced SIAR method to incorporate the effect of control, enabling it to utilize side-information constraints inherent to game-theoretic applications to model agents' responses to incentives from scarce data. MPC then leverages this model to implement dynamic incentive adjustments. Our experiments demonstrate the effectiveness of SIAR-MPC in guiding systems towards socially optimal equilibria, stabilizing chaotic and cycling behaviors. Notably, it achieves these results in data-scarce settings of few learning samples, where well-known system identification methods paired with MPC show less effective results.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (23)
  1. S. Smale, “Dynamics in general equilibrium theory,” The American Economic Review, vol. 66, no. 2, pp. 288–294, May 1976.
  2. E. J. Akin and V. Losert, “Evolutionary dynamics of zero-sum games,” Journal of Mathematical Biology, vol. 20, no. 3, pp. 231–258, Oct. 1984.
  3. Y. Sato, E. Akiyama, and J. D. Farmer, “Chaos in learning a simple two-person game,” Proceedings of the National Academy of Sciences, vol. 99, no. 7, pp. 4748–4751, Apr. 2002.
  4. S. Hart and A. Mas-Colell, “Uncoupled dynamics do not lead to Nash equilibrium,” The American Economic Review, vol. 93, no. 5, pp. 1830–1836, Dec. 2003.
  5. I. Sakos, A. Varvitsiotis, and G. Piliouras, “Discovering how agents learn using few data,” Jul. 2023.
  6. J. Riehl, P. Ramazi, and M. Cao, “A survey on the analysis and control of evolutionary matrix games,” Annual Reviews in Control, vol. 45, pp. 87–106, 2018.
  7. B. H. Zhang, G. Farina, I. Anagnostides, F. Cacciamani, S. M. McAleer, A. A. Haupt, A. Celli, N. Gatti, V. Conitzer, and T. Sandholm, “Steering no-regret learners to a desired equilibrium,” 2024.
  8. K. Paarporn, C. Eksin, J. S. Weitz, and Y. Wardi, “Optimal control policies for evolutionary dynamics with environmental feedback,” in 2018 IEEE Conference on Decision and Control (CDC), 2018, pp. 1905–1910.
  9. L. Gong, W. Yao, J. Gao, and M. Cao, “Limit cycles analysis and control of evolutionary game dynamics with environmental feedback,” Automatica, vol. 145, p. 110536, 2022.
  10. N. C. Martins, J. Certório, and R. J. La, “Epidemic population games and evolutionary dynamics,” Automatica, vol. 153, p. 111016, 2023.
  11. M. Raissi, P. Perdikaris, and G. Karniadakis, “Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations,” Journal of Computational Physics, vol. 378, pp. 686–707, 2019.
  12. Z. Chen, Y. Liu, and H. Sun, “Physics-informed learning of governing equations from scarce data,” Nature Communications, vol. 12, 2020.
  13. J. Stiasny, G. S. Misyris, and S. Chatzivasileiadis, “Physics-informed neural networks for non-linear system identification for power system dynamics,” in 2021 IEEE Madrid PowerTech, 2021, pp. 1–6.
  14. A. A. Ahmadi and B. E. Khadir, “Learning dynamical systems with side information,” SIAM Review, vol. 65, no. 1, pp. 183–223, 2023.
  15. M. Nagumo, “Über die Lage der Integralkurven gewöhnlicher Differentialgleichungen,” Proceedings of the Physico-Mathematical Society of Japan, vol. 24, pp. 551–559, 1942.
  16. P. A. Parrilo, “Structured semidefinite programs and semialgebraic geometry methods in robustness and optimization,” Ph.D. thesis, California Institute of Technology, California, Pasadena, United States, May 2000.
  17. J. B. Lasserre, “Global optimization with polynomials and the problem of moments,” SIAM Journal on Optimization, vol. 11, no. 3, pp. 796–817, Jan. 2001.
  18. P. A. Parrilo, “Semidefinite programming relaxations for semialgebraic problems,” Mathematical Programming, vol. 96, no. 2, pp. 293–320, May 2003.
  19. J. B. Lasserre, “A sum of squares approximation of nonnegative polynomials,” SIAM Journal on Optimization, vol. 16, no. 3, pp. 751–765, Jan. 2006.
  20. M. Laurent, “Sums of squares, moment matrices and optimization over polynomials,” in Emerging applications of algebraic geometry, 1st ed., ser. The IMA volumes in mathematics and its applications, M. Putinar, S. Sullivant, D. N. Arnold, and A. Scheel, Eds.   New York City, New York, United States: Springer, Sep. 2008, no. 149, pp. 157–270.
  21. E. Kaiser, J. N. Kutz, and S. L. Brunton, “Sparse identification of nonlinear dynamics for model predictive control in the low-data limit,” Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences, vol. 474, no. 2219, p. 20180335, 2018.
  22. S. Sorin, “No-regret algorithms in on-line learning, games and convex optimization,” Mathematical Programming, vol. 203, no. 1-2, pp. 645–686, Jan. 2024.
  23. W. Hu, G. Zhang, H. Tian, and Z. Wang, “Chaotic dynamics in asymmetric rock-paper-scissors games,” IEEE Access, vol. 7, pp. 175 614–175 621, 2019.
Citations (2)

Summary

No one has generated a summary of this paper yet.

Paper to Video (Beta)

No one has generated a video about this paper yet.

Whiteboard

No one has generated a whiteboard explanation for this paper yet.

Open Problems

We haven't generated a list of open problems mentioned in this paper yet.

Continue Learning

We haven't generated follow-up questions for this paper yet.

Collections

Sign up for free to add this paper to one or more collections.

Tweets

Sign up for free to view the 2 tweets with 0 likes about this paper.