Abstract
The purpose of this work is the comparison of learning algorithms in continuous time used in optimization and game theory. The first three are issued from no-regret dynamics and cover in particular “Replicator dynamics” and “Local projection dynamics”. Then we study “Conditional gradient” versus “Global projection” dynamics and finally “Frank-Wolfe” versus “Best reply” dynamics. Important similarities occur when considering potential or dissipative games.
Similar content being viewed by others
References
Akin E (1979) The geometry of population genetics. Lecture notes in biomathematics, vol 31. Springer, Berlin
Alvarez F, Bolte J, Brahic O (2004) Hessian Riemannian gradient flows in convex programming. SIAM J Control Optim 43:477–501
Antipin AS (1994) Minimization of convex functions on convex sets by means of differential equations. Differ Equ 30:1365–1375
Attouch H, Teboulle M (2004) Regularized Lotka-Volterra dynamical system as continuous proximal-like method in optimization. J Optim Theory Appl 121:541–570
Avrachenkov K, Borkar VS (2019) Metastability in stochastic replicator dynamics. Dyn Games Appl 9:366–390
Barron EN, Goebel R, Jensen RR (2010) Best response dynamics for continuous games. Proc AMS 138:1069–1083
Beck A, Teboulle M (2003) Mirror descent and nonlinear projected subgradient methods for convex optimization. Oper Res Lett 31:167–175
Benaim M, Hofbauer J, Sorin S (2005) Stochastic approximations and differential inclusions. SIAM J Control Optim 44:328–348
Benaim M, Hofbauer J, Sorin S (2006) Stochastic approximations and differential inclusions. Part II: applications. Math Oper Res 31:673–695
Benaim M, Hofbauer J, Sorin S (2012) Perturbations of set-valued dynamical systems, with applications to game theory. Dyn Games Appl 2:195–205
Bolte J (2003) Continuous gradient projection method in Hilbert spaces. J Optim Theory Appl 119:235–259
Bolte J, Teboulle M (2003) Barrier operators and associated gradient-like dynamical systems for constrained minimization problems. SIAM J Control Optim 42:1266–1292
Brézis H (1973) Opérateurs maximaux monotones et semi-groupes de contractions dans les espaces de Hilbert. North Holland, Amsterdam
Brown GW (1951) Iterative solutions of games by fictitious play. In: Koopmans TC (ed) Activity analysis of production and allocation. Wiley, Hoboken, pp 374–376
Brown GW, von Neumann J (1950) Solutions of games by differential equations. In: Kuhn HW, Tucker AW (eds) Contributions to the theory of games, I. Annals of mathematical studies, vol 24. Princeton University Press, Princeton, pp 73–79
Bruck RE (1975) Asymptotic convergence of nonlinear contraction semigroups in Hilbert space. J Funct Anal 18:15–26
Dafermos SC (1980) Traffic equilibrium and variational inequalities. Transp Sci 14:42–54
Dupuis P, Nagurney A (1993) Dynamical systems and variational inequalities. Ann Oper Res 44:9–42
Facchinei F, Pang J (2007) Finite-dimensional variational inequalities and complementarity problems. Springer, Berlin
Foster D, Young HP (1990) Stochastic evolutionary game dynamics. Theor Popul Biol 38:219–232
Frank M, Wolfe P (1956) An algorithm for quadratic programming. Naval Res Logist Q 3:95–110
Friesz TL, Bernstein D, Mehta NJ, Tobin RL, Ganjalizadeh S (1994) Day-to-day dynamic network disequilibria and idealized traveler information systems. Oper Res 42:1120–1136
Gilboa I, Matsui A (1991) Social stability and equilibrium. Econometrica 59:859–867
Hart S, Mas-Colell A (2003) Uncoupled dynamics do not lead to Nash equilibrium. Am Econ Rev 93:1830–1836
Hofbauer J, Sandholm WH (2009) Stable games and their dynamics. J Econ Theory 144:1665–1693
Hofbauer J, Sigmund K (1998) Evolutionary games and population dynamics. Cambridge University Press, Cambridge
Hofbauer J, Sorin S (2006) Best response dynamics for continuous zero-sum games. Discrete Contin Dyn Syst Ser B 6:215–224
Hofbauer J, Sorin S, Viossat Y (2009) Time average replicator and best reply dynamics. Math Oper Res 34:263–269
Kinderlehrer D, Stampacchia G (1980) An introduction to variational inequalities and their applications. Academic Press, London
Kwon J, Mertikopoulos P (2017) A continuous time approach to on-line optimization. J Dyn Games 4:125–148
Lahkar R, Sandholm WH (2008) The projection dynamic and the geometry of population games. Games Econ Behav 64:565–590
Maynard Smith J (1982) Evolution and the theory of games. Cambridge U.P., Cambridge
Mazumdar E, Ratliff LJ, Sastry SS (2020) On gradient-based learning in continuous games. SIAM J Math Data Sci 2:103–131
Mertikopoulos P, Sandholm WH (2016) Learning in games via reinforcement and regularization. Math Oper Res 41:1297–1324
Mertikopoulos P, Sandholm WH (2018) Riemannian game dynamics. J Econ Theory 177:315–364
Mertikopoulos P, Zhou Z (2019) Learning in games with continuous action sets and unknown payoff functions. Math Program 173:465–507
Minty GJ (1967) On the generalization of a direct method of the calculus of variations. Bull AMS 73:315–321
Monderer D, Shapley LS (1996) Potential games. Games Econ Behav 14:124–143
Monderer D, Shapley LS (1996) Fictitious Play property for games with identical interests. J Econ Theory 68:258–265
Moreau JJ (1965) Proximité et dualité dans un espace hilbertien. Bull Soc Math Fr 93:273–299
Nash J (1950) Equilibrium points in \(n\)-person games. Proc Natl Acad Sci 36:48–49
Nash J (1951) Non-cooperative games. Ann Math 54:286–295
Nemirovski A, Yudin D (1983) Problem complexity and method efficiency in optimization. Wiley, Hoboken
Nesterov Y (2009) Primal-dual subgradient methods for convex problems. Math Program 120:221–259
Nikaido H, Isoda K (1955) Note on non cooperative convex games. Pac J Math 5:807–815
Opial Z (1967) Weak Convergence of the sequence of successive approximations for nonexpansive mappings. Bull Am Math Soc 73:591–597
Pappalardo M, Passacantando M (2004) Gap functions and Lyapunov functions. J Glob Optim 28:379–385
Polyak B (1987) Introduction to optimization. In: Optimization software
Rockafellar RT (1970) Convex analysis. Princeton University Press, Princeton
Rockafellar RT (1970) Monotone operators associated with saddle-functions and minmax problems. In: Browder F (ed) Nonlinear functional analysis. Proceedings of symposia in pure math, vol 18. AMS, pp 241–250
Rosen JB (1965) Existence and uniqueness of equilibrium points for concave N-person games. Econometrica 33:520–534
Rustichini A (1999) Optimal properties of stimulus-response learning models. Games Econ Behav 29:230–244
Sandholm WH (2001) Potential games with continuous player sets. J Econ Theory 97:81–108
Sandholm WH (2011) Population games and evolutionary dynamics. MIT Press, Cambridge
Sandholm WH (2015) Population games and deterministic evolutionary dynamics. In: Young HP, Zamir S (eds) Handbook of game theory IV. Elsevier, Amsterdam, pp 703–778
Sandholm WH, Dokumaci E, Lahkar R (2008) The projection dynamic and the replicator dynamic. Games Econ Behav 64:666–683
Shahshahani S (1979) A new mathematical framework for the study of linkage and selection. In: Memoirs of the American Mathematical Society, vol 211
Smith MJ (1979) The existence, uniqueness and stability of traffic equilibria. Transp Res Part B 13:295–304
Sorin S (2009) Exponential weight algorithm in continuous time. Math Program Ser B 116:513–528
Sorin S (2011) On some global and unilateral adaptive dynamics. In: Sigmund K (ed) Evolutionary game dynamics. Proceedings of symposia in applied mathematics, vol 69. A.M.S., pp 81–109
Sorin S (2020) Replicator dynamics: old and new. J Dyn Games 7:365–385
Sorin S (2021) No-regret algorithms in on-line learning, games and convex optimization. In: Mathematical programming (to appear)
Sorin S, Wang C (2016) Finite composite games: equilibria and dynamics. J Dyn Games 3:101–120
Swinkels JM (1993) Adjustment dynamics and rational play in games. Games Econ Behav 5:455–484
Taylor PB, Jonker LB (1978) Evolutionary stable strategies and game dynamics. Math Biosci 40:145–156
Tsakas E, Voorneveld M (2009) The target projection dynamic. Games Econ Behav 67:708–719
Viossat Y (2014) Game dynamics and Nash equilibria. J Dyn Games 1:537–553
Wardrop G (1952) Some theoretical aspects of road traffic research communication networks. Proc Inst Civ Eng Part 2 1:325–378
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
I thank K. Avrachenkov, J. Bolte and J. Hofbauer for interesting discussions and nice comments. I acknowledge partial support from COST Action GAMENET.
This article is part of the topical collection “Multi-agent Dynamic Decision Making and Learning” edited by Konstantin Avrachenkov, Vivek S. Borkar and U. Jayakrishnan Nair.
Appendix
Appendix
Assume g continuous and dissipative and recall that X is convex and compact. Let us prove that \(S_{\mathrm{ext}} \) is non-empty. Define :
so that \( S_{\mathrm{ext}} = \cap _{y \in X} S_{\mathrm{ext}}^y\). Hence by compactness (weak-compactness in an Hilbert framework) it is enough to establish the following:
Claim
For any finite collection \(y_i \in X, i \in I\), there exists \( x \in co \{ y_i, i \in I\}\) such that:
Consider the finite two-person zero-sum game defined by the following \(I \times I\) matrix A:
Introduce \(B = \frac{1}{2}[ A +\, ^tA] \) and \(C = \frac{1}{2}[ A -\, ^tA] \).
The crucial point is that B has non negative coefficients since:
Hence an optimal strategy \(u \in \Delta (I)\) in the game C (which has value 0) gives \(uA_j \ge 0, \forall j \in I\). Letting \(x = \sum _i u_i y_i\) this writes as (47).
Rights and permissions
About this article
Cite this article
Sorin, S. Continuous Time Learning Algorithms in Optimization and Game Theory. Dyn Games Appl 13, 3–24 (2023). https://doi.org/10.1007/s13235-021-00423-x
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s13235-021-00423-x