Correlated equilibrium payoffs and public signalling in absorbing games

An absorbing game is a repeated game where some action combinations are absorbing, in the sense that whenever they are played, there is a positive probability that the game terminates, and the players receive some terminal payoff at every future stage.  We prove that every multi-player absorbing game admits a correlated equilibrium payoff. In other words, for every ε>0 there exists a probability distribution p _ε over the space of pure strategy profiles that satisfies the following. With probability at least 1−ε, if a pure strategy profile is chosen according to p _ε and each player is informed of his pure strategy, no player can profit more than ε in any sufficiently long game by deviating from the recommended strategy. Received: April 2001/Revised: June 4, 2002     

Stay-in-a-set games

There exists a Nash equilibrium (ε-Nash equilibrium) for every n-person stochastic game with a finite (countable) state space and finite action sets for the players if the payoff to each player i is one when the process of states remains in a given set of states G _i and is zero otherwise. Received: December 2000     

Value and perfection in stochastic games

A stochastic game isvalued if for every playerk there is a functionr ^k:S→R from the state spaceS to the real numbers such that for every ε>0 there is an ε equilibrium such that with probability at least 1−ε no states is reached where the future expected payoff for any playerk differs fromr ^k(s) by more than ε. We call a stochastic gamenormal if the state space is at most countable, there are finitely many players, at every state every player has only finitely many actions, and the payoffs are uniformly bounded and Borel measurable as functions on the histories of play. We demonstrate an example of a recursive two-person non-zero-sum normal stochastic game with only three non-absorbing states and limit average payoffs that is not valued (but does have ε equilibria for every positive ε). In this respect two-person non-zero-sum stochastic games are very different from their zero-sum varieties. N. Vieille proved that all such non-zero-sum games with finitely many states have an ε equilibrium for every positive ε, and our example shows that any proof of this result must be qualitatively different from the existence proofs for zero-sum games. To show that our example is not valued we need that the existence of ε equilibria for all positive ε implies a “perfection” property. Should there exist a normal stochastic game without an ε equilibrium for some ε>0, this perfection property may be useful for demonstrating this fact. Furthermore, our example sews some doubt concerning the existence of ε equilibria for two-person non-zero-sum recursive normal stochastic games with countably many states. This research was supported financially by the German Science Foundation (Deutsche Forschungsgemeinschaft) and the Center for High Performance Computing (Technical University, Dresden). The author thanks Ulrich Krengel and Heinrich Hering for their support of his habilitation at the University of Goettingen, of which this paper is a part.     

N-Person stochastic games with upper semi-continuous payoffs

Every n-person stochastic game with a countable state space, finite action sets for the players and bounded, upper semi-continuous payoffs has an ε-equilibrium for every ε>0.     

Van der waerden and ramsey type games

Let us consider the following 2-player game, calledvan der Waerden game. The players alternately pick previously unpicked integers of the interval {1, 2, ...,N}. The first player wins if he has selected all members of ann-term arithmetic progression. LetW*(n) be the least integerN so that the first player has a winning strategy. By theRamsey game on k-tuples we shall mean a 2-player game where the players alternately pick previously unpicked elements of the completek-uniform hypergraph ofN verticesK _N ^k , and the first player wins if he has selected allk-tuples of ann-set. LetR _k*(n) be the least integerN so that the first player has a winning strategy. We prove (W* (n))¹/n → 2,R ₂*(n)<(2+ε) ⁿ andR _k ^* n<2 ^nk / ^k! fork ≧3.     

A generalization of a result of R. Lyons about measures on [0, 1)

Letμ be a probability measure on [0, 1), invariant underS:x ↦px mod 1, and for which almost every ergodic component has positive entropy. Ifq is a real number greater than 1 for which logq/ logp is irrational, andT _n sendsx toq ⁿx mod 1, then for any ε>0 the measureμT _n ⁻¹ will — for a set ofn of positive lower density — be within ε of Lebesgue measure.     

Subsequence generators for ergodic group translations

LetT be an ergodic translation on a compact abelian group. For every infinite set of integers {n _i} and ε >0 there is a setA of measure less than ε such that {T ⁿ _iA} generates the σ-algebra of measurable sets. Research partially supported by National Science Foundation Grants MCS7703659 (first author) and MCS7606735A01 (second author).     

Stochastic games with non-observable actions

We examine n-player stochastic games. These are dynamic games where a play evolves in stages along a finite set of states; at each stage players independently have to choose actions in the present state and these choices determine a stage payoff to each player as well as a transition to a new state where actions have to be chosen at the next stage. For each player the infinite sequence of his stage payoffs is evaluated by taking the limiting average. Normally stochastic games are examined under the condition of full monitoring, i.e. at any stage each player observes the present state and the actions chosen by all players. This paper is a first attempt towards understanding under what circumstances equilibria could exist in n-player stochastic games without full monitoring. We demonstrate the non-existence of -equilibria in n-player stochastic games, with respect to the average reward, when at each stage each player is able to observe the present state, his own action, his own payoff, and the payoffs of the other players, but is unable to observe the actions of them. For this purpose, we present and examine a counterexample with 3 players. If we further drop the assumption that the players can observe the payoffs of the others, then counterexamples already exist in games with only 2 players.     

Limitations to Fréchet’s metric embedding method

Fréchet’s classical isometric embedding argument has evolved to become a major tool in the study of metric spaces. An important example of a Fréchet embedding is Bourgain's embedding [4]. The authors have recently shown [2] that for every ε>0, anyn-point metric space contains a subset of size at leastn ^1−ε which embeds into ℓ₂ with distortion . The embedding used in [2] is non-Fréchet, and the purpose of this note is to show that this is not coincidental. Specifically, for every ε>0, we construct arbitrarily largen-point metric spaces, such that the distortion of any Fréchet embedding into ℓ_p on subsets of size at leastn ^1/2+ε is Ω((logn)^1/p ). Supported in part by a grant from the Israeli National Science Foundation. Supported in part by a grant from the Israeli National Science Foundation. Supported in part by the Landau Center.     

Large Monotone Paths in Graphs with Bounded Degree

 We prove that for every ε>0 and positive integer r, there exists Δ₀=Δ₀(ε) such that if Δ>Δ₀ and n>n(Δ,ε,r) then there exists a packing of K _n with ⌊(n−1)/Δ⌋ graphs, each having maximum degree at most Δ and girth at least r, where at most εn ² edges are unpacked. This result is used to prove the following: Let f be an assignment of real numbers to the edges of a graph G. Let α(G,f) denote the maximum length of a monotone simple path of G with respect to f. Let α(G) be the minimum of α(G,f), ranging over all possible assignments. Now let α_Δ be the maximum of α(G) ranging over all graphs with maximum degree at most Δ. We prove that Δ+1≥α_Δ≥Δ(1−o(1)). This extends some results of Graham and Kleitman [6] and of Calderbank et al. [4] who considered α(K _n ). Received: March 15, 1999?Final version received: October 22, 1999     

The game for the speed of convergence in repeated games of incomplete information

We consider an infinitely repeated two-person zero-sum game with incomplete information on one side, in which the maximizer is the (more) informed player. Such games have value v _∞ (p) for all 0≤p≤1. The informed player can guarantee that all along the game the average payoff per stage will be greater than or equal to v _∞ (p) (and will converge from above to v _∞ (p) if the minimizer plays optimally). Thus there is a conflict of interest between the two players as to the speed of convergence of the average payoffs-to the value v _∞ (p). In the context of such repeated games, we define a game for the speed of convergence, denoted SG _∞ (p), and a value for this game. We prove that the value exists for games with the highest error term, i.e., games in which v _n (p)− v _∞ (p) is of the order of magnitude of . In that case the value of SG _∞ (p) is of the order of magnitude of . We then show a class of games for which the value does not exist. Given any infinite martingale 𝔛^∞={X _k }^∞ _k=1, one defines for each n : V _n (𝔛^∞) ≔E∑ⁿ _k=1 |X _k+1 − X _k|. For our first result we prove that for a uniformly bounded, infinite martingale 𝔛^∞, V _n (𝔛^∞) can be of the order of magnitude of n ^1/2−ε, for arbitrarily small ε>0. Received January 1999/Final version April 2002     

On the existence of automorphisms with simple Lebesgue spectrum

It is shown that ifT is a measure preserving automorphism on a probability space (Ω,B, m) which admits a random variable X₀ with mean zero such that the stochastic sequence X₀ o T_n,n ε ℤ is orthonormal and spans L₀ ²(Ω,B,m), then for any integerk ≠ 0, the random variablesX o T^nk,n ε ℤ generateB modulom.     

On The Structure Of Triangle-Free Graphs Of Large Minimum Degree

It is shown that for every ε>0 there exists a constant L such that every triangle-free graph on n vertices with minimum degree at least (1/3+ε)n is homomorphic to a triangle-free graph on at most L vertices. * Research partially supported by KBN grant 2 P03A 016 23.     

Quitting games – An example

Quitting games are multi-player sequential games in which, at every stage, each player has the choice between continuing and quitting. The game ends as soon as at least one player chooses to quit; each player i then receives a payoff r _S ⁱ, which depends on the set S of players that did choose to quit. If the game never ends, the payoff to each player is zero.? We exhibit a four-player quitting game, where the “simplest” equilibrium is periodic with period two. We argue that this implies that all known methods to prove existence of an equilibrium payoff in multi-player stochastic games are therefore bound to fail in general, and provide some geometric intuition for this phenomenon. Received: October 2001     

Ideal einstein,conformally flat and semi-symmetric immersions

Recently, B. Y. Chen introduced a new intrinsic invariant of a manifold, and proved that everyn-dimensional submanifold of real space formsR ^m (ε) of constant sectional curvature ε satisfies a basic inequality δ(n ₁,…,n _k )≤c(n ₁,…,n _k )H ²+b(n ₁,…,n _k )ε, whereH is the mean curvature of the immersion, andc(n ₁,…,n _k ) andb(n ₁,…,n _k ) are constants depending only onn ₁,…,n _k ,n andk. The immersion is calledideal if it satisfies the equality case of the above inequality identically for somek-tuple (n ₁,…,n _k ). In this paper, we first prove that every ideal Einstein immersion satisfyingn≥n ₁+…+n _k +1 is totally geodesic, and that every ideal conformally flat immersion satisfyingn≥n ₁+…+n _k +2 andk≥2 is also totally geodesic. Secondly we completely classify all ideal semi-symmetric hypersurfaces in real space forms. The author was supported by the NSFC and RFDP.     

A problem of Erdős on abelian groups

The following theorem is proved. Ifa ₁,a ₂, ...a _n are nonzero elements inZ _n , and are not all equal, then ε₁ a ₁+ε₂ a ₂+...+ε _n a _n =0 has at leastn solutions with ε _i =0 or 1.     

Mesure de Hausdorff des complementaires desU ε de zygmund

A. Zygmund originally introduced the notion of sets of unicityU _ε, and showed that they differ from classicalU-sets in that they can have positive measure. He then asked if they could be of full measure. J. P. Kahame and Y. Katznelson proved recently that there wereU _ε of full measure. The object of this paper is to show that, in terms of Hausdorff measure, one cannot go beyond that result, for a general sequence ε. In the case of a given sequence ε, and a given Hausdorff determining functionh, it gives a simple test for determining the existence ofU _ε with a complement of zero Hausdorff measure. In this paper the proof of the main known results concerning the measure ofU _ε sets is also reproduced.      

Nearly flat almost monotone measures are big pieces of lipschitz graphs

A concentrated (ξ, m) almost monotone measure inR ⁿ is a Radon measure Φ satisfying the two following conditions: (1) Θ ^m (Φ,x)≥1 for every x ∈spt (Φ) and (2) for everyx ∈R ⁿ the ratioexp [ξ(r)]r^−mΦ(B(x,r)) is increasing as a function of r>0. Here ξ is an increasing function such thatlim _r→0-ξ(r)=0. We prove that there is a relatively open dense setReg (Φ) ∋spt (Φ) such that at each x∈Reg(Φ) the support of Φ has the following regularity property: given ε>0 and λ>0 there is an m dimensional spaceW ⊂R ⁿ and a λ-Lipschitz function f from x+W into x+W^‖ so that (100-ε)% ofspt(Φ) ∩B (x, r) coincides with the graph of f, at some scale r>0 depending on x, ε, and λ.