Ezio Marchi
Director of Instituto de Matemática Aplicada San Luis, Universidad Nacional de San Luis, Ejército de los Andes 950, San Luis, Argentina
Correspondence to: Ezio Marchi, Director of Instituto de Matemática Aplicada San Luis, Universidad Nacional de San Luis, Ejército de los Andes 950, San Luis, Argentina.
Email: | |
Copyright © 2024 The Author(s). Published by Scientific & Academic Publishing.
This work is licensed under the Creative Commons Attribution International License (CC BY).
http://creativecommons.org/licenses/by/4.0/
Abstract
In game theory there are many discussions about the concept of rationality. Theoretically these discussions are concerned with an appropriate mathematical concept which derives from the concept itself. In this paper we present a new way of looking rationality which is concerned with an external view of the rules of games. We prove the existence of these new friendly equilibrium and perturbed friendly equilibria.
Keywords:
Equilibria, Friendly
Cite this paper: Ezio Marchi, Friendly Equilibria, Journal of Game Theory, Vol. 13 No. 2, 2024, pp. 32-36. doi: 10.5923/j.jgt.20241302.02.
1. Results
The most important equilibrium of non-cooperative games is the Nash equilibrium point; the perfect equilibrium of Selten and proper equilibrium of Myerson are also considered important. Most authors of non-cooperative games study applications of these three equilibria. However, there was no plan to consider friends. In this work we emphasize this last concept that can be applied to problems of sociology, economics, mathematics, etc. In the future it can be thought that the existence of this friendly equilibria can be applied to the aforementioned balances. We prove the existence.Here we are going to introduce the general tools of the game. Consider an n-person game in normal formwhere the strategy sets are non-empty and finite. The mixed extension isNow for a player consider a finite sequence of playerswhich we call the “successor friends” of player . Then from an intuitive point of view it is natural to assume that the player is “rational”: whenever his payoff is the same he is going to act in such way that first he will maximize the payoff of his first successor friend . In the case that the payoff of this successor friend is equal again he will act to maximize the payoff of his second successor friend and so on until the last .Consider for the setsFormally we define a friendly equilibrium point if .Theorem 1: Given any game and is upper semi-continuous for each , there always exists a friendly equilibrium.Proof: given any point consider the set of pointssince the expectations are multi-linear then each is non-empty convex and compact. Moreover, by the continuity of the expected functions and the hypothesis such, are upper semi-continuous. Therefore using Kakutani’s fixed point applied to there exists a fixed point . Such point is a friendly equilibrium (q.e.d.).As a simple observation we would like to point out that in the case that for each our new notion coincides with the fundamental Nash equilibrium points.It is interesting to remark that the existence of a friendly equilibrium is equivalent to the existence of a solution of the non-linear systemwhere denotes the discrete support of . The proof of the equivalence is rather easy and is left to the reader.
2. Perturbed Friendly Equilibria
Now we will extend the concept of friendly equilibria to the theory of perturbed equilibria. For this, we will follow the approaches of Selten (1975) and Myerson (1978).Consider a point such that it satisfiesTheorem 2: A point is a friendly equilibrium point if and only if it fulfills B) with for each .Proof: We prove it by induction on . For the case for each , this is very well known. Assume that it is valid for , then we will show that it is also valid for .For this, consider that for each . Suppose that for and Let be the point defined as thenwhich is impossible since and is identified with .Inversely, consider a point such thatfor each , then only if andtherefore for each .Then following the ideas of Selten (1975), we define a perfect friendly equilibrium as follows: given , a point is called a -perfect friendly equilibria ifA perfect friendly equilibrium is a limit point of a sequence where is an -perfect friendly rational equilibria. Similarly following the ideas of Myerson (1978) we define -proper friendly equilibria asClearly an -proper friendly equilibrium is an -perfect friendly equilibrium. A point is called a proper friendly equilibria if is a limit point of a sequence where is an -proper friendly equilibrium.When the ’s corresponding concepts become the perfect and proper ones respectively.Theorem 3: Every normal form game possesses at least one proper friendly equilibrium if , is upper semi-continuous.Proof: In order to prove this existence theorem let us observe that for given positive integers it always holds true This can be proven by induction on , which we skip. Therefore, it is clear thatNow given , definewhere is the cardinality of , and letFor define the correspondence from by:We have that is closed and convex for every and the mapping is upper semi-continuous. Now we will prove that . Considerfor all and . Clearly Let the pointClearly .Now if then and | (1) |
| (2) |
since and the observation given above.Now if then all Then performing the same operation as before instead in, in, it is easy to see that Let be the n-tuple . Then it satisfies the condition of the Kakutani Fixed Point Theorem (Kakutani, 1941) and the fixed point of is an -proper friendly equilibrium. Now for each we have an -proper friendly equilibrium, thus making , it is always possible to find a converging subsequence since the space is compact. The limit point of such subsequence is a friendly equilibrium point. The proof is complete, then the theorem follows.As a simple consequence of this result we haveCorollary: Every normal form game has at least one perfect friendly equilibrium.Finally we would like to say that if the do not belong to and are given externally, we have a external new notion of equilibrium. And we would like to emphasize that the previous material might be generalized accordingly for the continuous case. In this situation the upper semi-continuity condition appears more natural.
ACKNOWLEDGMENTS
I am great full for the collaboration of Miss Gimena Páez in editing and taping this paper.
Note
1. This paper was written in the year 1990 in a project supported by CONICET.
References
[1] | Kakutani, S. (1941). A generalization of Brouwer fixed point theorem. Duke Math. J. 451- 459. |
[2] | Marchi, E. (1991-1992). -subgame perfect equilibrium. |
[3] | Myerson, R. B. (1978). Refinements of the Nash equilibrium concept. Int. Game Theory 7, 73-80. |
[4] | Selten, R. (1975). Reexamination of the perfectness concept for equilibrium points in extensive games. Int. J. Game Theory 4, 25-55. |
[5] | Van Damme (1987). Stability and perfection of Nash Equilibria. Springer-Verlag, Berlin – New York. |