International Journal of Analysis and Applications
ISSN 2291-8639
Volume 14, Number 1 (2017), 20-26
http://www.etamaths.com

ALGEBRAIC HYPER-STRUCTURES ASSOCIATED TO NASH EQUILIBRIUM

POINT AND APPLICATIONS

A. DELAVAR KHALAFI AND B. DAVVAZ∗

Abstract. In this paper, we generalize some concepts of the game theory such as Nash equilibrium

point, saddle point and existence theorems on hyper-structures. Based on new definitions and theo-

rems, we obtain some important results in the game theory. A few suitable examples have been given
for better understanding.

1. Introduction and preliminaries

Algebraic hyperstructures are suitable generalizations of classical algebraic structures. In a classical
algebraic structure, the composition of two elements is an element, while in an algebraic hyperstructure,
the composition of two elements is a set. More exactly, if H is a non-empty set and P∗(H) is the set of
all non-empty subsets of H, then we consider maps of the following type: fi : H×H −→P∗(H), where
i ∈ {1, 2, . . . ,n} and n is a positive integer. The maps fi are called (binary) hyperoperations. For all
x,y of H, fi(x,y) is called the (binary) hyperproduct of x and y. An algebraic system (H,f1, . . . ,fn)
is called a (binary) hyperstructure. Usually, n = 1 or n = 2. Under certain conditions, imposed to the
maps fi, we obtain the so-called semihypergroups, hypergroups, hyperrings or hyperfields. Sometimes,
external hyperoperations are considered, which are maps of the following type: h : R×H −→P∗(H),
where R 6= H. Usually, R is endowed with a ring or a hyperring structure. Several books have been
written on this topic, see [1, 2, 6, 13]. Hyperstructure theory both extends some well-known group
results and introduce new topics leading us to a wide variety of applications, as well as to a broadening
of the investigation fields, for example see [4, 5, 8, 10–12]. A recent book on hyperstructures [2] points
out on their applications in rough set theory, cryptography, codes, automata, probability, geometry,
lattices, binary relations, graphs and hypergraphs. Another book [6] is devoted especially to the study
of hyperring theory. Several kinds of hyperrings are introduced and analyzed. The volume ends with
an outline of applications in chemistry and physics, analyzing several special kinds of hyperstructures:
e-hyperstructures and transposition hypergroups. The theory of suitable modified hyperstructures can
serve as a mathematical background in the field of quantum communication systems.

Optimization theory is the study of the extremal values of a function: its minima and maxima. In
mathematics, optimization refers to choosing the best element from some set of available alternatives.
Nonlinear programming deals with the problem of optimizing an objective function in the presence of
some constraints. In [8], we generalized the optimization theory on algebraic hyperstructures. The
game theory is another framework which has been generalized of optimization theory. The famous
mathematician Von Numan has been proposed his important game theory in 1928. Game theory is an
important branch of applied mathematics in which decision maker chooses his strategy with regards
to strategies of other players. In this theory any player tries to choose his best strategy for obtaining
maximum pay off function. The methods and applications in the game theory are very different. In [9],
bileaner two person nonzero sum game has been considered.

In the mathematical domain the extensions of the previous works are very popular. Our contribution
in this paper is the extensions of some previous concepts in game theory on the hyperstructures and
have considered several new examples that we can’t solve them via usual game theory. It means that
we have obtained a new widespread in the game theory.

2010 Mathematics Subject Classification. 20N20, 91A99.

Key words and phrases. algebraic hyperstructure; optimization; game theory; Nash equilibrium point.

c©2017 Authors retain the copyrights of
their papers, and all open access articles are distributed under the terms of the Creative Commons Attribution License.

20



ALGEBRAIC HYPER-STRUCTURES ASSOCIATED TO NASH EQUILIBRIUM POINT AND APPLICATIONS 21

2. Game Theory

In this paper, we address a hyper-structures as follows:

? : H ×H → H ⊗H ⊆P∗(H), (2.1)
· : F ×H → H,

+ : H ×H → H,
where H 6= ∅, ? is a commutative hyperoperation such that ?(H×H) = H⊗H, · and + are commutative
binary operations and F is a filed. Henceforth, let F = R. Convex and concave functions play an
important role in almost all branches of mathematics as well as other areas of science and engineering.
Convex and concave functions have many special and important properties. In this paper we use some
these properties in game theory . Let define P∗(X) = {x ? y ∈ H ⊗H : x, y ∈ X} for all non-empty
subset X in H. Now, we formulate the n-person game theory problem on hyper-structures as follows:
Define the function

fi : P
∗(X1) ×···×P∗(Xn) → R, i = 1, · · · ,n,

where Xi is any non-empty subset of H and fi and Xi are a pay off function and a strategy set of
i-th player, respectively. Let W = X1 ×···×Xn. The next definition plies an important role in the
following discussions.

Definition 2.1. The n-tuple (x1 ?y1, · · · ,xn ?yn) ∈ P∗(X1)×···×P∗(Xn) is called Nash equilibrium
point, if the following inequalities hold.

f1(x1 ? y1, · · · ,xn ? yn) 6 f1(x1 ? y1, · · · ,xn ? yn), (2.2)
for all x1 ? y1 ∈ P∗(X1),

...

fn(x1 ? y1, · · · ,xn ? yn) 6 fn(x1 ? y1, · · · ,xn ? yn),
for all xn ? yn ∈ P∗(Xn).

Henceforth, for simplicity let consider n = 2. As a special case we consider the following situation.
If we have

f1(x1 ? y1,x2 ? y2) + f2(x1 ? y1,x2 ? y2) = 0,

for all (x1 ?y1,x1 ?y1) ∈ P∗(X1)×P∗(X2). Thus, the Nash equilibrium point satisfies in the following
inequalities,

f1(x1 ? y1,x2 ? y2) 6 f1(x1 ? y1,x2 ? y2) 6 f1(x1 ? y1,x2 ? y2), (2.3)

for all (x1 ? y1,x2 ? y2) ∈ P∗(X1) ×P∗(X2), or equivalently
f2(x1 ? y1,x2 ? y2) 6 f2(x1 ? y1,x2 ? y2) 6 f2(x1 ? y1,x2 ? y2), (2.4)

for all (x1 ? y1,x2 ? y2) ∈ P∗(X1) ×P∗(X2). The next definition plays an important role in the game
theory.

Definition 2.2. The pair (x1 ? y1,x2 ? y2) ∈ P∗(X1) ×P∗(X2) is called the saddle point, if satisfies
the first inequalities (3) or second inequalities (4).

Suppose that

v = inf
x2,y2∈X2

sup
x1,y1∈X1

f(x1 ? y1,x2 ? y2) and v = sup
x1,y1∈X1

inf
x2,y2∈X2

f(x1 ? y1,x2 ? y2).

Clearly, we have v ≤ v.

Definition 2.3. The strategy x1, y1 ∈ X1 is called max-min if
v = inf

x2,y2∈X2
f(x1 ? y1,x2 ? y2),

and similarly, the strategy x2 ? y2 ∈ X2 is called min-max, if
v = sup

x1,y1∈X1
f(x1 ? y1,x2 ? y2).

The following theorem gives us a necessary and sufficient condition that guaranties the existence of
saddle point.



22 DELAVAR KHALAFI AND DAVVAZ

Theorem 2.1. Suppose that the pay-off function f(x1 ? y1,x2 ? y2) on X1 ×X2 is given. There is a
saddle point (x1 ? y1,x2 ? y2) if and only if

sup
x1,y1∈X1

inf
x2,y2∈X2

f(x1 ? y1,x2 ? y2) = inf
x2,y2∈X2

sup
x1,y1∈X1

f(x1 ? y1,x2 ? y2). (2.5)

In addition x1 ? y1 ∈ P∗(X1) and x2 ? y2 ∈ P∗(X2) are max-min and min-max, respectively.

Proof. Suppose that (x1 ? y1,x2 ? y2) is a saddle point. We have

v ≤ sup
x1,y1∈X1

f(x1 ? y1,x2 ? y2) = f(x1 ? y1,x2 ? y2) = inf
x2,y2∈X2

f(x1 ? y1,x2 ? y2) ≤ v,

so the equation (5) is held. Now, suppose that we have the equation (5) and x1 ? y1 and x2 ? y2 are
max-min and min-max strategies, respectively. Then,

f(x1 ? y1,x2 ? y2) ≤ sup
x1,y1∈X1

f(x1 ? y1,x2 ? y2) = v

= v = inf
x2,y2∈X2

f(x1 ? y1,x2 ? y2)

≤ f(x1 ? y1,x2 ? y2).
This completes the proof. �

The saddle point does not always exist. The following example denotes such a situation.

Example 2.1. Suppose that H = R and X1 = X2 = [0, 1]. We define f : P∗([0, 1]) −→ R and ?
respectively as follows:

x ? y = x, f(x1 ? y1,x2 ? y2) = 3x
2
1 − 5x1x2 + 3x

2
2.

Clearly, M(x1 ? y1) = min
x2,y2∈X2

f(x1 ? y1,x2 ? y2) =
11x21
12

. Then, we have: v = max
x1,y1∈X1

11x21
12

= 11
12

,

x1 = 1. Similarly, N(x2 ? y2) = maxx1,y1∈X1 f(x1 ? y1,x2 ? y2) = max{3x22, 3x22 − 5x2 + 3}, x2 =
3
5

.

Therefore, we have v = min
x2,y2∈X2

N(x2 ? y2) =
27
25

, v < v. Now, we obtain max-min and min-max

strategies. As

min
x2,y2∈X2

f(1 ? y1,x2 ? y2) = min
x2∈X2

3 − 5x2 + 3x22 =
11

12
= v,

max
x1,y1∈X1

f(x1 ? y1,
3

5
? y2) = max

x1∈X1
3x21 − 3x1 +

27

25
=

27

25
= v.

Therefore, {1 ? y1 : y1 ∈ [0, 1]} and {35 ? y2 : y2 ∈ [0, 1]} are max-min and min-max strategy set,
respectively. According to Theorem (2.5), (1 ? y1,

3
5
? y2) is not a saddle point.

The above example shows that we must generalize the previous saddle point definition.

Definition 2.4. Let � > 0. The pair (x�1 ? y
�
1,x

�
2 ? y

�
2) ∈ P∗(X1) ×P∗(X2) is called �-saddle point of

f(x1 ? y1,x2 ? y2) on X1 ×X2, if
f(x1 ? y1,x

�
2 ? y

�
2) − � ≤ f(x

�
1 ? y

�
1,x

�
2 ? y

�
2) ≤ f(x

�
1 ? y

�
1,x2 ? y2) + � (2.6)

for all x1, y1 ∈ X1, x2, y2 ∈ X2.

Lemma 2.1. Let x1 ?y1, x2 ?y2 are max-min and min-max strategies, respectively and � = v−v ≥ 0.
Then, (x1 ? y1, x2 ? y2) is an �−saddle point of f(x1 ? y1,x2 ? y2) on X1 ×X2.

Proof. If v = v, so x1 ? y1, x2 ? y2 is a saddle point. Let � = v −v > 0. Clearly, x1 ? y1, x2 ? y2 is an
�−saddle point. �

Using the above definition, we consider the following example.

Example 2.2. In Example 2, let � = v −v. As
f(x1 ? y1,

3
5
? y2) − � = 3x21 − 3x1 +

27
25
− ( 27

25
− 11

12
)

≤ f(1 ? y1, 35 ? y2) =
27
25

≤ f(1 ? y1,x2 ? y2) + �

= 3 − 5x2 + 3x22 + (
27
25
− 11

12
),



ALGEBRAIC HYPER-STRUCTURES ASSOCIATED TO NASH EQUILIBRIUM POINT AND APPLICATIONS 23

for all x1, y1 ∈ X1, x2, y2 ∈ X2, so (x�1 ? y�1,x�2 ? y�2) = (1 ? y1,
3
5
? y2), is a �−saddle point of

f(x1 ? y1,x2 ? y2).

Similar to saddle point, we can generalize the concepts of min-max and max-min as follows:

Definition 2.5. Let � > 0. The strategies x�1 ? y
�
1 and x

�
2 ? y

�
2 are �− max-min and �−min-max, if

inf
x2,y2∈X2

f(x�1 ? y
�
1,x2 ? y2) ≥ v − � and sup

x1,y1∈X1
f(x1 ? y1,x

�
2 ? y

�
2) ≤ v + �.

In the remaining part, we consider the some new topology characters of hyper-structures [7] and
their applications in game theory. Let H be a metric space, X, Y be compact subsets of H and

Y (x1 ? y1) = Arg min
x2, y2∈X2

f(x1 ? y1,x2 ? y2)

= {x̂2 ? ŷ2|x̂2, ŷ2 ∈ X2, f(x1 ? y1, x̂2 ? ŷ2)

= min
x2, y2∈X2

f(x1 ? y1,x2 ? y2)}.

Theorem 2.2. Suppose that the pay off function f(x1?y1,x2?y2) is a continuous function on X1×X2
and P∗(X1), P

∗(X2) are compact sets in H ⊗ H. Then, the function g(x1 ? y1) = min
x2, y2∈X2

f(x1 ?

y1,x2 ? y2) on X1 is a continuous function.

Proof. Let {xk1} and {yk1} be two sequences in X1 such that convergence to x1 and y1, respectively.
By considering g(xk1 ? y

k
1 ), we can prove that it converges to g(x1 ? y1). In contradiction, there

are subsequences {xkl1 }, {y
kl
1 } in X1 such that lim

l→∞
g(x

kl
1 ? y

kl
1 ) 6= g(x1 ? y1). Choosing sequence

{xkl2 ?y
kl
2 ∈ Y (x

kl
1 ?y

kl
1 )}, based on compactness of X2, we have lim

l→∞
x
kl
2 ?y

kl
2 = x2 ?y2. We must show

that x2 ? y2 ∈ Y (x1 ? y1). By definition xkl2 ? y
kl
2 , we have

g(x
kl
1 ? y

kl
1 ) = f(x

kl
1 ? y

kl
1 ,x

kl
2 ? y

kl
2 ) ≤ f(x

kl
1 ? y

kl
1 ,x2 ? y2),

for all x2, y2 ∈ X2. In the above inequality, when l →∞, we conclude that
f(x1 ? y1,x2 ? y2) ≤ f(x1 ? y1,x2 ? y2),

for all x2, y2 ∈ X2. Therefore, x2 ? y2 ∈ Y (x1 ? y1), that is

lim
l→∞

g(x
kl
1 ? y

kl
1 ) = lim

l→∞
f(x

kl
1 ? y

kl
1 ,x

kl
2 ? y

kl
2 ) = f(x1 ? y1,x2 ? y2) = g(x1 ? y1).

�

Under what conditions is the function Y (x1 ? y1) continuous? We assert the sufficient condition,
that guaranties the continuous function Y (x1 ? y1).

Theorem 2.3. Suppose that the conditions of previous theorem have been satisfied and for any x1, y1 ∈
X1, Y (x1 ? y1) = {x2 ? y2} be a singleton. Then, Y (x1 ? y1) is a continuous function on X1.

Proof. Suppose that the function Y (x1 ?y1) isn’t continuous in x1 ?y1 ∈ P∗(X1), so there is a sequence
{xk1 ? yk1} in P

∗(X1), such that it converges {x1 ? y1}, but {xk1 ? yk1} = {x
k
2 ? y

k
2} does not convergent

to Y (x1 ?y1) = {x2 ?y2}. Therefore, there is an �−Neighborhood N?� (Y (x1 ?y1)) ⊂ P∗(X2) that does
not contain infinite number of elements {xk2 ? yk2}. Rely on compactness of P∗(X2) −N?� (Y (x1 ? y1)),
we have subsequence {Y (xkl1 ?y

kl
1 )} = {x

kl
2 ?y

kl
2 }⊂ P

∗(X2) −N?� (Y (x1 ?y1)), such that it convergent
to x′2 ? y

′
2 6= x2 ? y2. According to the previous theorem, we conclude that x′2 ? y′2 = Y (x1 ? y1), which

it contradicts to singleton. �

Definition 2.6. Let f : P∗(X) → R, where X is non-empty convex subset in H. The function f is
called a convex function on P∗(X) if

f([λx1 + (1 −λ)x2] ? [λy1 + (1 −λ)y2]) ≤ λf(x1 ? y1) + (1 −λ)f(x2 ? y2)
for each x1,x2,y1,y2 ∈ X, x1 ? y1, x2 ? y2 ∈ P∗(X) and for all 0 ≤ λ ≤ 1. The function is
called strictly convex on P∗(X) if the inequality is satisfied as a strict inequality for each distinct
x1 ? y1, x2 ? y2 ∈ P∗(X) and 0 < λ < 1. The function f is called concave (strictly concave) on X if
−f is convex (strictly convex) on X.



24 DELAVAR KHALAFI AND DAVVAZ

The following function is an example of convex function [8].

Example 2.3. Let H = R+ ×R+. Suppose that

zmin = min{x1,x2,y1,y2},
zmax = max{x1,x2,y1,y2},
(x1,y1) ? (x2,y2) = [zmin, zmax] × [zmin, zmax] ⊆ R+ ×R+

and f : H⊗H → R is defined by f((x1,y1)?(x2,y2)) = zmax−zmin, for all (x1,y1), (x2,y2) ∈ X, where
X is any non-empty convex subset in H. Suppose that ((x̄1, ȳ1), (x̄2, ȳ2), z̄) and ((x̂1, ŷ1), (x̂2, ŷ2), ẑ) in
epi?f and x̄λ = λx̄1 + (1 −λ)x̄2, ȳλ = λȳ1 + (1 −λ)ȳ2, x̂λ = λx̂1 + (1 −λ)x̂2, ŷλ = λŷ1 + (1 −λ)ŷ2.
One can show that

max{x̄λ, ȳλ, x̂λ, ŷλ}− min{x̄λ, ȳλ, x̂λ, ŷλ}≤ λz̄ + (1 −λ)ẑ.

That is, ((x̄λ, ȳλ), (x̂λ, ŷλ), λz̄ + (1 −λ)ẑ) ∈ epi?f. Therefore, epi?f is a convex set, so we conclude
that the function f is also a convex function.

Theorem 2.4. Let H = Rn, X1, X2 be convex and compact subsets of H and f(x1 ? y1,x2 ? y2) is
a continuous function on X1 × X2. Suppose that for any x2 ? y2 ∈ P∗(X2), f(x1 ? y1,x2 ? y2) be a
concave function with respect to x1 ?y1 ∈ P∗(X1) and for any x1 ?y1 ∈ P∗(X1) it is a convex function
with respect to x2 ? y2 ∈ P∗(X2). Then, the function f(x1 ? y1,x2 ? y2) has a saddle point.

Proof. : At first , we consider the special case that the function f(x1 ?y1,x2 ?y2) be a strictly convex
w.r.t x2 ? y2 ∈ P∗(X2). Then, for any x1 ? y1 ∈ P∗(X1), the function f(x1 ? y1,x2 ? y2) obtains its
unique minimum on X2 in Y (x1 ? y1). According to previous theorems, we conclude g(x1 ? y1) and
Y (x1 ? y1) are continuous function on X1. Suppose that the function g(x1 ? y1) obtains its minimum
on X1 in x1 ? y1. We can show that (x1 ? y1,Y (x1 ? y1)) is a saddle point of f(x1 ? y1,x2 ? y2) on
X1×X2. Let x1 ?y1 ∈ P∗(X1) be arbitrary, 0 < λ < 1 and Y = Y ([(1−λ)x1 +λx1]?[(1−λ)y1 +λy1]).
According to concavity of the function f(x1 ? y1,x2 ? y2) with respect to x1 ? y1 ∈ P∗(X1), we have

(1 −λ)g(x1 ? y1) + λf(x1 ? y1,Y ([(1 −λ)x1 + λx1] ? [(1 −λ)y1 + λy1]))
≤ (1 −λ)f(x1 ? y1,Y ([(1 −λ)x1 + λx1] ? [(1 −λ)y1 + λy1]))

+λf(x1 ? y1,Y ([(1 −λ)x1 + λx1] ? [(1 −λ)y1 + λy1]))
≤ f([(1 −λ)x1 + λx1] ? [(1 −λ)y1 + λy1],Y ([(1 −λ)x1 + λx1] ? [(1 −λ)y1 + λy1]))
= g([(1 −λ)x1 + λx1] ? [(1 −λ)y1 + λy1])
≤ g(x1 ? y1).

Therefore, λf(x1 ?y1,Y ([(1−λ)x1 + λx1] ? [(1−λ)y1 + λy1])) ≤ λg(x1 ?y1). Divided by λ and it tends
to zero, we have the following inequalities:

f(x1 ? y1,Y (x1 ? y1)) ≤ g(x1 ? y1)
= f(x1 ? y1,Y (x1 ? y1))
≤ f(x1 ? y1,x2 ? y2),

for all x1, y1 ∈ X1 and x2, y2 ∈ X2. In general we consider the following perturbed function
f�(x1 ?y1,x2 ?y2) = f(x1 ?y1,x2 ?y2) + �h(x2 ?y2) where h(x2 ?y2) is a continuous and strictly convex
function on X2. Clearly f�(x1 ?y1,x2 ?y2) is a continuous, concave with respect to x1 ?y1 and strictly
convex with respect to x2 ? y2 function. Using previous discussion, we have

f�(x1 ? y1,x
�
2 ? y

�
2) ≤ f�(x

�
1 ? y

�
1,x

�
2 ? y

�
2) ≤ f�(x

�
1 ? y

�
1,x2 ? y2), (2.7)

for all x1, y1 ∈ X1 and x2, y2 ∈ X2. Let � = �k in the inequalities (7) and �k → 0+. Because of
compactness of X1 and X2, we conclude that x

�
1 ? y

�
1 → x1 ? y1 and x�2 ? y�2 → x2 ? y2. Therefore,

(x1 ? y1,x2 ? y2) is a saddle point of the function f(x1 ? y1,x2 ? y2). �

3. Applications

In this section, we consider some examples in game theory and explain our theory in the previous
section.



ALGEBRAIC HYPER-STRUCTURES ASSOCIATED TO NASH EQUILIBRIUM POINT AND APPLICATIONS 25

3.1. Examples. The following examples show that, how we can use Nash equilibrium point in practice.
First of all, we show that the Nash equilibrium point gives a generalization of usual optimization
problem.

Example 3.1. Since the optimization problem is a special case of the game theory with one player, so
we obtain the optimization model as follows:

f1(x1 ? y1) 6 f1(x1 ? y1),

for all x1 ? y1 ∈ P∗(X1). This means that x1 ? y1 ∈ Argmin{f(x1 ? y1) : x1 ? y1 ∈ P∗(X1)}.

Example 3.2. Let H = {f ∈ C|f : R → R} and X1, X2 ⊆ H, where X1 and X2 are such continuous
functions that have left and right inverse. We define the following function.

f1(g,h) =

∫
Dg

⋂
Dh

[(g −h)(x)]2dx,

where g and h are the left and right inverse of an arbitrary function and Dg and Dh are the domains
of g and h, respectively. Let f2(g,h) = −f1(g,h). Clearly (g,g) = (f,f) is a saddle point, i.e., the left
and right inverse are the same.

3.2. Numerical Examples.

Example 3.3. In Example 3, we define X1 = {(x, 0) : 0 ≤ x ≤ M}, X2 = {(0,y) : 0 ≤ y ≤
N}. Clearly, X1, X2 are compact and convex subsets of H. According to Example 3, the function
h(x1?y1) = zmax−zmin = max{x11,x21} and e(x2?y2) = zmax−zmin = max{y12,y22}, where x1 = (x11, 0),
y1 = (x

2
1, 0), x2 = (0,y

1
2), y2 = (0,y

2
2), x1, y1 ∈ X1, x2, y2 ∈ X2 are convex functions. Let

f(x1 ? y1, x2 ? y2) = e(x2 ? y2) −h(x1 ? y1).
We can show that the function f(x1 ? y1,x2 ? y2) is a continuous, concave function with respect to
x1?y1) ∈ P∗(X1) and convex function with respect to x2?y2) ∈ P∗(X2). Therefore, the all assumptions
of Theorem (2.11) has been held. Clearly, {(x1 ? y1,x2 ? y2) : max{x11,x21} = max{y12,y22}} is the set
of saddle points of the function f(x1 ? y1,x2 ? y2) on X1 ×X2.

Example 3.4. Let define H = R and X1 = X2 = H, x ? y = [min{x,y},max{x,y}],
f(x1 ? y1,x2 ? y2) = (x

2
2 + y

2
2) − (x

2
1 + y

2
1),

for all x1, y1 ∈ X1 and x2, y2 ∈ X2. In this example, X1, X2 are not compact subset of H. We
choose x1 = y1 = x2 = y2 = 0. Then, x1 ? y1 = {0}, x2 ? y2 = {0} and f({0},{0}) = 0. Therefore,
(x1 ? y1,x2 ? y2) = ({0},{0}) is a saddle point, that is

f(x1 ? y1,x2 ? y2) = −(x21 + y
2
1) ≤ f(x1 ? y1,x2 ? y2) = 0 ≤ f(x1 ? y1,x2 ? y2) = (x

2
2 + y

2
2),

for all x1, y1 ∈ X1 and x2, y2 ∈ X2. Thus, Theorem (2.11) is a sufficient condition.

References

[1] P. Corsini, Prolegomena of Hypergroup Theory, Second edition, Aviani Editore, 1993.

[2] P. Corsini and V. Leoreanu, Applications of Hyperstructure Theory, Advances in Mathematics, Kluwer Academic
Publishers, Dordrecht, 2003.

[3] I. Cristea and M. Stefanescu, Hypergroups and n-ary relations, Eur. J. Comb. 31 (2010), 780-789.

[4] B. Davvaz, A. Dehghan Nezhad and A. Benvidi, Chemical hyperalgebra: Dismutation reactions, MATCH Commun.
Math. Comput. Chem. 67 (2012) 55-63.

[5] B. Davvaz, A. Dehghan Nezhad and A. Benvidi, Chain reactions as experimental examples of ternary algebraic
hyperstructures, MATCH Commun. Math. Comput. Chem. 65 (2011) 491-499.

[6] B. Davvaz and V. Leoreanu-Fotea, Hyperring Theory and Applications, International Academic Press, USA, 2007.

[7] A. Dehghan Nezhad and B. Davvaz, Universal hyperdynamical systems, Bull. Korean Math. Soc. 47 (2010), 513-526.
[8] A. Delavar Khalafi and B.Davvaz, Algebraic hyper-structures associated to convex analysis and applications, Filomat

26 (2012) 55-65.

[9] A. Delavarkhalafi Solution methods for equilibrium points in bileaner non-zero sum games, Ph.D. Thesis (2002) (In
Russian).

[10] M. Ghadiri, B. Davvaz and R. Nekouian, Hv-Semigroup structure on F2-offspring of a gene pool, Int. J. Biomath.

5 (2012), Art. ID 1250011.
[11] S. Hošková and J. Chvalina, Discrete transformation hypergroups and transformation hypergroups with phase

tolerance space, Discre. Math. 308 (2008), 4133-4143.



26 DELAVAR KHALAFI AND DAVVAZ

[12] V. Leoreanu-Fotea, B. Davvaz, Fuzzy hyperrings, Fuzzy Sets and Syst. 160 (2009), 2366-2378.
[13] T. Vougiouklis, Hyperstructures and Their Representations, Hadronic Press, Florida, 1994.

Department of Mathematics, Yazd University, Yazd, Iran

∗Corresponding author: davvaz@yazd.ac.ir


	1. Introduction and preliminaries
	2. Game Theory
	3. Applications
	3.1. Examples
	3.2.  Numerical Examples

	References