D. Gale, F.M. Stewart. Infinite Games with Perfect Information, in H. W. Kuhn, A. W. Tucker eds. Contributions to the Theory of Games, Volume II, pp. 245-266, Annals of Mathematics Studies, Princeton University Press, 1953.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....# has a winning strategy in Gn(#(G) Proof. cf. 16, theorem 11] or [15, theorem 14.13] Write A for #(G) Trivially, # has a winning strategy in G0 (A) Assume now that n 0, and suppose, for contradiction, that # has no winning strategy in Gn(A) Since finite length games are determined [9], #must have a winning strategy in this game. By proposition 5.3, for some set W of 2 elements of A, # has a winning strategy that only ever directs him to choose elements in W. W =W # a ; b : a, b #W # A. Then W 2 2 4 4 . Since n # 1, we have 2 2 # 4 , so W 5 4 . ....

D. Gale and F. Stewart, Infinite games with perfect information, Contributions to the theory of games II, Annals of mathematical studies 28 (1953), 291--296.

....is that if player B wins this new game, sometimes in the first game B could not have possibly played optimally . It is well known that AC (the axiom of choice) and AD are contradictory. The usual proof uses a diagonal argument combined with the fact that the number of strategies is 2 0 [1], 2] The status of AD has been examined in great depth [4] 5] There seems to be two approaches. One can accept AC and ask which sets are determined. This leads to questions which are independent of the usual axiomatization of set theory [2] 4] 5] The other and more radical approach is to ....

D.Gale, F.Steward ; Infinite games with perfect information, in Contributions to the Theory of Games II Ann. Math. Studies 28 (1953) p 245-266

....the tree of positions. Player I s aim is to get this path into a previously fixed winning set while player II tries to THE CATEGORY OF INNER MODELS 259 prevent this. The winning set is determined if there is a winning strategy for one of the players. By a classical result of GALE and STEWART [GS53] topologically simple winning sets are determined. We shall show that winning sets representable by an ENFW are also determined. This is done by introducing an auxiliary game G which is an extension of the original game G by side moves . One can view the original game as the auxiliary game ....

David Gale, Frank M. Stewart, Infinite games with perfect information, in: Harold W. Kuhn, Albert W. Tucker (eds.), Contributions to the Theory of Games II, Princeton 1953 [Annals of Mathematical Studies 28], p. 245--266

....the game continues from the position (OE; X[F; n] In the beginning, the position is (OE; f;g) The above game is a game of perfect information: the strategies of both players are allowed to depend on the whole sequence of previous positions. Hence it is determined by the Gale Stewart theorem [2], as a zero sum game of perfect information. For example, if A has at least 2 elements, player I has a winning strategy in G(A;8v 0 9v 1 =v 0 (v 0 = v 1 ) The winning strategy consists of player I doing nothing, player II simply has only losing moves. Theorem 6 A j= OE if and only if player II ....

....game ends. Player II is the winner, if A j= Xn OE ) B j= Yn OE holds for all atomic and negated atomic formulas OE. Otherwise player I wins. This is also a game of perfect information and the concept of winning strategy is defined as usual. The game is determined by the Gale Stewart theorem ([2]) 8 Lemma 8 Suppose A is an infinite model and B a finite model of the empty vocabulary. Then player I has a winning strategy in EF 4 . Proof. The first move of player I uses Case 3 with V = He chooses an element a 2 A and the function F : f;g A defined by F ( a. Suppose II answers ....

David Gale and Frank Stewart. Infinite games with perfect information, In "Contributions to the theory of games II", Ann. Math. Studies 28 (1953), 245-266.

....mainly while visiting Institute for Advanced Study, Princeton, USA, and supported by a grant of the state of New Jersey. Supported by the grant A1019901 of the Academy of Sciences of the Czech Republic. 1 models in which such principles fail. We consider countable games, a concept introduced in [1], and which plays an important role in contemporary set theory. We consider two principles. Both are based on playing simultaneously several games. We call a strategy for playing several games a parallel strategy. The first principle says, roughly speaking, that when playing the same game ....

D. Gale, F. Steward, Infinite games with perfect information, Ann. Math. Studies 28 (1953), 245-266

....paths as Path(T) To indicate when an agent W wins, it is enough to identify the set of all such paths in Path(T) say W, called the winning set for W, where we assign the victory to W. Within the theory of two player games, it is usually assumed that the winning sets of players are complementary (Gale Stewart 1953). Such games are called win loose games. For general multi agent graph games it is usually not the case (e.g. in chess no win situation is possible) Thus, in our games each agent A i is associated with its winning set W i . We call the list (A 0 , W 0 ; A m 1 , W m 1 ) the combined winning ....

Gale, D., Stewart, F. M. (1953) Infinite Games with Perfect Information, Contributions to the theory of games, Ann. of Math. Studies, No. 28, Princeton Univ. Press, pp. 245266, 1953.

....(note that for all i the target belongs to B i ) We know that CAP (B ; 1, thus there exists a point c 1 which is at distance at most from any point in B , and in particular from t. For the other direction, suppose that Player I does not have a winning strategy, it follows (see e.g. [13]) that Player II has a winning strategy. Let the teacher regard the student as a simulator for Player I, and choose the r i s according to the winning strategy of Player II. Note that, in any step, the set B i is the set of all possible targets consistent with the answers given during the first ....

Gale D., and F. M. Stewart, "Infinite Games with Perfect Information", Annals of Mathematics, Vol. 28, pp. 245-266, 1953. 19

....#) and the result is a play of # . # won by the proponent. 5 Determinacy A classical theorem of Zermelo [27] see also [11]) asserts that if a game (of perfect information, between two players, in which every play is a win for exactly one player) is such that it always ends in a finite number of moves, then the game is determined in the sense that one of the players has a winning strategy. This theorem poses a serious ....

....Zermelo s theorem is to violate the hypothesis of that theorem by allowing games to have infinitely long plays. This is the approach used in [4] Its feasibility rests on the existence of undetermined games when infinite plays are allowed; such games were constructed (using the axiom of choice) in [11]. Here the criterion for winning can no longer be merely successful completion of the protocol (lest the game be determined again, by another result from [11] but must be given as part of the data of the game. In fact, to get an undetermined game, the criterion for winning must be quite ....

[Article contains additional citation context not shown here]

David Gale and F. M. Stewart. Infinite games with perfect information. Ann. Math. Studies, 28:245--266, 1953.

....all formal models of Discrete Event Systems, but holds for the Supervisory Control Theory Framework we adopted. 14 of the scenario is to investigate the indefinite, nonterminating interaction between the two agents. The problem can be also formulated in a game theoretic setting as introduced by [GS53] In this case the interaction between two agents is viewed as an infinite two player game, where the two players choose alternate actions from alphabets Sigma c and Sigma d , denoting the alphabet of the controller and the disturber. In the supervisory control theory [RW89] setting the two ....

D. Gale and F. M. Stewart. Infinite games with perfect information, chapter in Contributions to the theory of games. Princeton University Press, 1953.

....to a winning strategy for 9 in the game G n ( A) In order to express OE n as a universal formula more explicitly, let us consider what these winning strategies are. First, since G n is of finite length it is a determined game, meaning that if 9 does not have a winning strategy then 8 does [Gale and Stewart, 1953]. So what is a strategy for 8 It is a set of instructions telling him how to play each round of the game and the instructions for the ith round of the game (0 i n) must depend only on player 9 s previous moves, i.e. on a sequence p = p 0 ; p i Gamma1 ) 2 fIn; Outg i . Let us write ....

D Gale and F Stewart. Infinite games with perfect information. Contributions to the theory of games II, Annals of mathematical studies, 28:291--296, 1953.

....to a winning strategy for 9 in the game Gn ( A) In order to express OE n as a universal formula more explicitly, let us consider what these winning strategies are. First, since Gn is of finite length it is a determined game, meaning that if 9 does not have a winning strategy then 8 does [GS53] So what is a strategy for 8 It is a set of instructions telling him how to play each round of the game and the instructions for the ith round of the game (0 i n) must depend only on player 9 s previous moves, i.e. on a sequence p = p 0 ; p i Gamma1 ) 2 fIn; Outg i . Let us ....

D Gale and F Stewart. Infinite games with perfect information. Contributions to the theory of games II, Annals of mathematical studies, 28:291--296, 1953.

....motivated by standard ideas from logic, was used, for example, in [2] It works well in the context of infinite games (as in [2] but seems seriously deficient when games are required to terminate after a finite number of moves. The reason is that, by a classical theorem of Gale and Stewart [5], such finitely long games always admit winning strategies for one or the other player. This means that, for any protocol A, either there is a behavior (a winning strategy in the sense mentioned above) for the server in A or there is a winning strategy for the client, i.e. a strategy by which the ....

D. Gale and F. M. Stewart, Infinite games with perfect information, Ann. Math. Studies 28 (1953), 245--266.

....an algorithm, is not feasible. We adopt the following game theoretic framework (see [Tho95] for more detailed background) The interaction between two parties (say, a program and its environment) is modeled by two players, called 0 and 1 here, of an infinite game (or Gale Stewart game [GS53]) In a play of the game, both parties perform actions in turn, thus building up an infinite computation. In the state based description to be assumed in this paper, an action causes a transition in a state graph, and referring to the alternation between the two players, we suppose that this state ....

D. Gale and F. M. Stewart. Infinite games with perfect information. Annals of Mathematical Studies, 28:245 -- 266, 1953.

....of the initial papers concerning competitive analysis. Alternatively, Raghavan and Snir [16] formulate the concept of competitiveness in terms of infinite games. They develop analogues of our Theorem 2. 1 (using classical results concerning determinacy in infinite games see Gale and Stewart [9], Martin [13] and Theorem 2.2. Raghavan and Snir [16] discuss the relation between these two approaches; in particular, they give a sufficient condition for when the alternative definitions of competitiveness are equivalent. 2 Definitions and Results We study the performance of online ....

D. Gale and F.M. Stewart. Infinite games with perfect information. In W.H. Kuhn and A.W. tucker, Editors, Contributions to the Theory of Games Vol. II, Annals of Mathematics Studies, 28, pages 245--266. Princeton University Press, Princeton, New Jersey, 1953.

....you to beat us It is remarkable that ordinary computation can find this strategy when presented with the game in question (McNaughton, 1993) No ordinary computer can literally play the game, of course. However, for a game utterly beyond the Turing Limit, see the undetermined game featured in (Gale and Stewart, 1953): this is a game where a winning strategy cannot be devised by ordinary computation (in fact, there is no mathematical function which is a winning strategy ) It seems to us that infinite games, perhaps especially uncomputable infinite games, provide promising frameworks for mindmachine ....

D. Gale and F.M. Stewart. Infinite games with perfect information. In Annals of Math Studies 28-- Contributions to the Theory of Games, pages 245-- 266. Princeton University Press, Princeton, NJ, 1953.

.... Gamma(A; W ) is a game with complete information, but do not worry about the definition of games with complete information: we will not use it. And yes, determinacy is problematic in the infinite case. An example of an indeterminate infinite game appeared already in the paper by Gale and Stewart [6] where they introduced infinite games of interest to us here. Actually, Gale and Stewart rediscovered infinite games in the context of emerging game theory. Similar infinite games were studied by East European mathematicians much earlier [11] Here is their example. Let A be the arena of binary ....

....the example with users competing for a printer, consider a particular request R of a printer. It is easy to see that the collection of branches, i.e. computations, where R is satisfied is open and gives rise to an open game. And I am happy to prove the determinacy of open games for you. Theorem 1 [6] Every closed or open game is determinate. ffl Q: Isn t enough to speak about open games only ffl A: Well, there is this little asymmetry: Player 1 begins. Let me start with a few definitions. A support for an open set U of branches of an arena A is any subset S of A such that S x2S [x] U ....

D. Gale and F.M. Stewart. Infinite Games with Perfect Information. Ann. of Math. Studies 28 (Contributions to the Theory of Games II), 245--266. Princeton, 1953.

....For some payoff functions f , such as bounded Borel measurable functions f , it has been proven that the infinite game of perfect information Gamma p:i: f) is determined. But using the Axiom of Choice, a nonmeasurable payoff function f can be constructed such that Gamma(f ) is not determined [10]. The axiom AD, the axiom that all games Gamma(f ) are determined, is widely used as an alternative to AC [9, 11] A game of infinite duration and imperfect information is similar, except that both players make their n th move at the same time. These games are called Blackwell games, named ....

D. Gale and F.M. Stewart, Infinite Games with Perfect Information, in: Contributions to the theory of Games, Annals of Mathematics Studies 28, p.245-266, Princeton University Press, Princeton, N.J., 1953.

No context found.

D. Gale, F.M. Stewart. Infinite Games with Perfect Information, in H. W. Kuhn, A. W. Tucker eds. Contributions to the Theory of Games, Volume II, pp. 245-266, Annals of Mathematics Studies, Princeton University Press, 1953.

No context found.

D. Gale and F. Stewart, Infinite games with perfect information, Contributions to the theory of games II, Annals of mathematical studies 28 (1953), 291--296.

No context found.

D. Gale and F. Stewart (1953). Infinite games with perfect information. Ann. Math. Studies, 28:245-266.

No context found.

D. Gale & F. Stewart, "Infinite games with perfect information," Ann. Math. Stud. 28 (1953), 245 -- 266; MR 14,999b.

No context found.

David Gale, Frank M. Stewart, Infinite Games with Perfect Information, in: Harold W. Kuhn, Albert W. Tucker (eds.), Contributions to the Theory of Games II, Princeton 1953 [Annals of Mathematical Studies 28], p. 245--266

No context found.

D. Gale and F.M. Stewart. Infinite games with perfect information. Contributions to the Theory of Games, pages 245--266, 1953. 18

No context found.

GaS D. Gale, F. M. Stewart, Infinite games with perfect information, in: Contributions to the theory of games, II, Ann. Math. Studies 28 (1953) pp. 245--266.

No context found.

D. Gale and F. M. Stewart [1953], Infinite games with perfect information, Contributions to the Theory of Games, Ann. of Math. Stud. 2(28), 245--266, Princeton.

First 50 documents

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC