Volume 2013, Issue 1 952021
Research Article
Open Access

Existence Solutions of Vector Equilibrium Problems and Fixed Point of Multivalued Mappings

Kanokwan Sitthithakerngkiet

Kanokwan Sitthithakerngkiet

Department of Mathematics, Faculty of Science, Naresuan University, Phitsanulok 65000, Thailand nu.ac.th

Search for more papers by this author
Somyot Plubtieng

Corresponding Author

Somyot Plubtieng

Department of Mathematics, Faculty of Science, Naresuan University, Phitsanulok 65000, Thailand nu.ac.th

Search for more papers by this author
First published: 04 February 2013
Academic Editor: Qamrul Hasan Ansari

Abstract

Let K be a nonempty compact convex subset of a topological vector space. In this paper-sufficient conditions are given for the existence of xK such that F(T)∩VEP(F) ≠ ∅, where F(T) is the set of all fixed points of the multivalued mapping T and VEP(F) is the set of all solutions for vector equilibrium problem of the vector-valued mapping F. This leads us to generalize and improve some existence results in the recent references.

1. Introduction

Let X be a Hausdorff topological vector space, a fixed point of such a multi-valued mapping T : X → 2X, where 2X denotes, the family of subsets of X means a point p in X such that pT(p) and the set of all fixed points of T is denoted by F(T). The study of fixed point theorems of multi-valued mappings started from von Neumann [1] in case of continuous mappings to multi-valued mappings. Since then, various notions of the fixed point theorems for the multi-valued mappings have been studied in [24]. Recently, the fixed point theorems for multi-valued mapping were generalized and improved by many authors: see, for example, [510].

On the other hand, the equilibrium problems were introduced by Blum and Oettli [11] and by Noor and Oettli [12] in 1994 as generalizations of variational inequalities and optimization problems. The equilibrium problem theory provides a novel and united treatment of a wide class of problems which arise in economics, finance, image reconstruction, ecology, transportation, network, elasticity, and optimization. This theory has had a great impact and influence in the development of several branches of pure and applied sciences. Classical examples of equilibrium problems are variational inequalities, optimization problems, and complementarity problems. Presently, many results on the existence of solutions for vector variational inequalities (in short, VVI) and vector equilibrium problems (in short, VEP) have been established (see, e.g., [1320]). For the more generalized from of (VEP) and (VVI) as special case, we assume that X and Y are Hausdorff topological vector spaces, K is a nonempty convex subset of X, and C is a pointed closed convex cone in Y with int C. Let T : K → 2K be a set-valued mapping and for a given vector valued mapping F : K × KY such that F(x, x) = 0 for each xK, the vector quasi-equilibrium problem (VQEP) and the vector quasi-variational inequality (VQVI), respectively: find xK such that
()
where A : KL(X, Y), and L(X, Y) is denoted by the space of all continuous linear operators for X to Y: see, for example, [2126] and references therein.

From our mentions on the importance of the fixed point theorems and the equilibrium problems, we have an inspired question to find when the set of the solution of both problems will have a joint solution, not an empty set.

To answer the question, we assume that T : K → 2K is a set-valued mapping and for a given vector-valued mapping F : K × KY such that F(x, x) = 0 for each xK, let us present the fixed point problem of multi-valued mapping together with the vector equilibrium problem; in particular, it is to find xK such that
()
This problem shows the relationship in sense of intersection between fixed points of the multi-valued mapping and the vector equilibrium problem so the set of all solutions of the problem (2) is denoted by F(T)∩VEP(F). This problem includes vector quasi-equilibrium problems (in short, VQEP) and vector quasi-variational inequalities (in short, VQVI) as special cases.

The main purpose of this paper, we provide sufficient conditions and prove the existence solutions of intersection between the set of all fixed points of the multi-valued mapping and the set of all solutions for vector equilibrium problem by using the generalization of the Fan-Browder fixed point theorem. We also study the existence solutions of intersection between the set of all fixed points of the multi-valued mapping and the set of all solutions for vector variational inequality. Consequently, our results extend the existence theorems of vector quasi-equilibrium problems and vector quasi-variational inequalities.

2. Preliminaries

Throughout this paper, unless otherwise specified, we assume that X, and Y are Hausdorff topological vector spaces, K is a nonempty convex subset of X and C is a pointed closed convex cone in Y with int C. From problem (2), some special cases are as follows.

(I) If T(x) ≡ K for all xK, then problem (2) reduces to the vector equilibrium problem (in short, VEP): find xK such that
()
The set of all solutions for vector equilibrium problem is denoted by VEP(F).
Moreover, if we set Y = and C = [0, ), then VEP reduces to the equilibrium problem that the set of all solutions is denoted by EP (F): find xK such that
()
(II) If Y = and C = [0, ), then problem (2) reduces to the problem: find xK such that
()
Moreover, this problem includes the quasi-equilibrium problems (in short, QEPs) considered and studied by Lin and Park [27] which is to find xK such that
()
(III) If F(x, y) =   Ax, yx〉 for all x, yK, problem (3) reduces to the vector variational inequality (in short, VVI): find xK such that
()
The solution sets of problem (7) is denoted by VVI(K, A), where A : KL(X, Y), and L(X, Y) is denoted by the space of all continuous linear operators for X to Y.
Furthermore, if Y = and C = [0, ), then VVI reduces to the variational inequality that the set of all solutions is denoted by VI (K, A): find xK such that
()
(IV) If F(x, y) =   Ax, yx〉 for all x, yK, then problem (2) reduces to the following problems that is to find xK such that
()
This problem shows the relationship in sense of intersection between fixed points of the multi-valued mapping and the vector variational inequality so the set of all solutions of the problem (9) is denoted by F(T)∩VVI(K, A).
Let us recall some concepts and properties that are needed in this sequel. Given the multi-valued mappping T : X → 2Y, the inverse T−1 of T is the multi-valued map from (T), the range of T, to X defined by
()
The mapping f : XY is continuous at some point xX if and only if for any neighborhood V of f(x), there is a neighborhood U of x such that f(U)⊆V.

Definition 1 (see [28].)Let X and Y be topological vector spaces. Let K be a nonempty subset of X and C be a point closed convex cone in Y with int  C, where int  C denotes the topological interior of C. A bifunction F : K × KY is said to be C-strongly pseudomonotone if, for any x, yK,

()
A mapping f : KY is said to be C-convex if for all x, yK and for all λ ∈ [0,1],
()
And the mapping f is said to be hemicontinuous if, for all x, yK and for all λ ∈ [0,1],
()

Remark 2. If Y = and C = [0, ), then

  • (1)

    the C-strongly pseudomonotonicity of F : K × KY reduces to the monotonicity of F : K × K (i.e., F(x, y) + F(y, x) ≤ 0 for all x, yK). In fact, F(x, y)∉−C∖{0}⇔F(x, y) ≥ 0, this implies that −F(y, x) ≥ 0 and it is equivalence to F(y, x)∈−C;

  • (2)

    the C-convexity of f : KY reduces to the convexity of f : K (i.e., f(λx + (1 − λ)y) ≤ λf(x)+(1 − λ)f(y)).

Definition 3 (see [29].)Let X be a topological space and let Y be a set. A map T : X → 2Y is said to have the local intersection property if for each xX with T(x) ≠ there exists an open neighborhood V(x) of x such that ⋂zV(x)T(z) ≠ .

The following lemma is useful in what follows and can be found in [30].

Lemma 4. Let X be a topological space and let Y be a set. Let T : X → 2Y be a map with nonempty values. Then, the following are equivalent.

  • (i)

    T has the local intersection property.

  • (ii)

    There exists a map F : X → 2Y s.t. F(x) ⊂ T(x) for each xX, F−1(y) is open for each yY and X = ⋃yYF−1(y).

Subsequently, Browder [2] obtained in 1986 the following fixed point theorem.

Theorem 5 (Fan-Browder fixed point theorem). Let X be a nonempty compact convex subset of a Hausdorff topological vector space and T : X → 2X be a map with nonempty convex values and open fibers (i.e., for yY, T−1(y) is called the fiber of T on y). Then T has a fixed point.

The generalization of the Fan-Browder fixed point theorem was obtained by Balaj and Muresan [31] in 2005 as follows.

Theorem 6. Let X be a compact convex subset of a t.v.s and T : X → 2X be a map with nonempty convex values having the local intersection property. Then T has a fixed point.

3. Main Theorem

In this section, the existence solutions of the fixed point for multi-valued mappings and the vector equilibrium problems will be presented. To do this, the following lemma is necessary.

Lemma 7. Let K be a nonempty and convex subset of X. Let T : K → 2K be a set-valued mapping such that for any xK, T(x) is nonempty convex subset of K. Assume that F : K × KY be a hemicontinuous in the first argument, C-convex in the second argument, and C-strong pseudomonotone. Then the following statements are equivalent.

  • (i)

    Find xK such that xT(x) and F(x, y) ∉ −C∖{0}  yK.

  • (ii)

    Find xK such that xT(x) and F(y, x) ∈ −C  yK.

Proof. (i)→(ii) It is clear by the C-strong pseudomonotone.

(ii)→(i) Let xK such that

()
For any yK and α ∈ (0,1), we set zα = αy + (1 − α)x and so we have zαK because K is convex. By the assumption, we conclude that
()
Since F is C-convex in the second argument and by (15), we get
()
This implies that αF(zα, y) ∈ C and since C is a convex cone then we have F(zα, y) ∈ C. Since F is a hemicontinuous in the first argument and zαx as α → 0+, we have F(x, y) ∈ C for all yK. Therefore we obtain that
()
This completes the proof.

Theorem 8. Let K be a nonempty compact convex subset of X and let F : K × KY be a C-strong pseudomonotone, hemicontinuous in the first argument and C-convex, l.s.c. in the second argument such that 0 = F(x, x) for all xK. Let T : K → 2K be a set-valued mapping such that for any xK, T(x) is nonempty convex subset of K and for any yK, T−1(y) is open in K. Assume the set P : = {xXxT(x)} is open in K and for any xK, T(x)∩{yK   |   F(y, x)∉−C} ≠ . Then F(T)∩ VEP (F) ≠ .

Proof. For any xK, we define the set-valued mapping A, B : K → 2K by

()
Also, we define the set-valued mapping H : K → 2K by
()
Then we have H(x) is convex. Indeed, let y1, y2B(x) and α ∈ (0,1). Since F is C-convex in the second argument, we have
()
Then αy1 + (1 − α)y2B(x) and hence B(x) is convex. Since T(x) is convex, then H(x) is also convex.

By the defining of H, we see that H has no fixed point Indeed, suppose that there is xK such that xH(x). It is impossible for xKP, then xP and so xB(x). Thus F(x, x)∈−C∖{0}, a contradiction with 0 = F(x, x). Using the contrapositive of Theorem 6, we obtain that H has no local intersection property. Define the set-valued mapping G : K → 2K by

()
From the C-strong pseudmonotonicity of F, we have G(x)⊆H(x) for any xK. Next, we will show that for each yK, G−1(y) is open in K. For any yK, we denote the complement of A−1(y) by [A−1(y)] C = {xKF(y, x)∈−C}. Since C is closed and F is l.s.c. in the second argument, we have [A−1(y)] C is closed in K and so A−1(y) is open in K. We note that
()
Since for any yK, T−1(y), A−1(y), and P are open, we have G−1(y) is open in K. Thus, by the contrapositive of Lemma 4, we have
()
Hence, there exists such that for all yK. That is . If then A(x)∩T(x) = , which contradicts with the assumption. Therefore, and A(x) = . This implies that and F(y, x)∈−C for all yK. This completes the proof by Lemma 7.

The following example guarantees the assumption that the set T(x)∩A(x) ≠ , where A(x) = {yKF(y, x) ∉ −C}.

Example 9. Let X, Y = , K = [−1,1], and C = [0, ). For any x, yK, we define two mappings F : K × K → 2Y and T : K → 2K by

()
Clearly, T(x) is nonempty convex subset of K and T−1(y) is open in K. If F(x, y) ∉ −C∖{0} for all x, yK, then xy and it implies that for xy, F(x, y)∈−C. This shows that F is C-strong pseudomonotone. Let x, y1, y2K and α ∈ [0,1] and since 0 ∈ K, we obtain that
()
Then F is C-convex in the second argument and it is easy to see that F is a hemicontinuous in the first argument and l.s.c. in the second argument.

Note that

()
If −1 ≤ x ≤ 0, then T(x) = (−1 − x, 1] which includes (0,1]. Also (0,1] is contained A(x) for all −1 ≤ x ≤ 0. Otherwise, (x, 1]⊆(x − 1,1] for any 0 ≤ x < 1. This is to confirm the set T(x)∩A(x) ≠ for each xK (see Figure 1). Moreover, this example asserts that the set P = {xXxT(x)} is open in K because it is equal to the set (−0.5,1] which is open in K.

Description unavailable

Taking T(x) ≡ K for all xK in Theorem 8, we have the following results.

Corollary 10. Let K be a nonempty compact convex subset of X and F : K × KY be a C-strong pseudomonotone, hemicontinuous in the first argument and C-convex, l.s.c. in the second argument such that 0 = F(x, x) for all xK. Then, VEP has a solution.

If we set the vector-valued mapping F ≡ 0, then Theorem 8 reduces to the following corollary introduced by Browder (see [2, Theorem 1]).

Corollary 11. Let K be a nonempty compact convex subset of X. Let T : K → 2K be a set-valued mapping such that for any xK, T(x) is a nonempty convex subset of K and for any yK, T−1(y) is open in K. Then there exists in K such that .

If we set Y = and C = [0, ) in Theorem 8 together with Remark 2, we have the following result.

Corollary 12. Let K be a nonempty compact convex subset of X and let F : K × K be a monotone, hemicontinuous in the first argument and convex, l.s.c. in the second argument such that 0 = F(x, x) for all xK. Let T : K → 2K be a set-valued mapping such that for any xK, T(x) is a nonempty convex subset of K and for any yK, T−1(y) is open in K. Assume the set P : = {xXxT(x)} is open in K and for any xK, T(x)∩{yKF(y, x) > 0} ≠ . Then F(T)∩EP (F) ≠ .

Let L(X, Y) be a space of all linear continuous operators from X to Y. A mapping A : KL(X, Y) is said to be C-strong pseudomonotone if it satisfies
()
and it is called hemicontinuous if, for all x, yK and for all λ ∈ [0,1], the mapping λ ↦ 〈T(x + λ(yx), z)〉 is continuous at 0+.

As a direct consequence of Theorem 8, we obtain the following result.

Theorem 13. Let K be a nonempty compact convex subset of X and let A : KL(X, Y) be a C-strong pseudomonotone and hemicontinuous. Let T : K → 2K be a set-valued mapping such that for any xK, T(x) is nonempty convex subset of K and for any yK, T−1(y) is open in K. Assume the set P : = {xX   |   xT(x)} is open in K and for any xK, T(x)∩{yK   |   A(y), xy〉∉−C} ≠ . Then there exists such that

()

Proof. We define the vector value mapping F : K × KY by

()
We will show that F satisfies all conditions in Theorem 8. Clearly F(x, x) = 0 and by the assumptions of A, we have F is C-strong pseudomonotone and hemicontinuous in the first argument. Let xK be fixed. For any y, zK and θ ∈ [0,1], we obtain that
()
Then F is C-convex in the second argument.

Next, we will show that F is l.s.c. in the second argument. Let xK be fixed. Let y0K and N be a neighborhood of F(x, y0). Since the linear operator Ax is continuous, there exists an open neighborhood M of y0 such that for all yM, 〈Ax, yx〉 ∈ N because N is a neighborhood of 〈Ax, y0x〉. Thus for all yM, F(x, y) ∈ N. Hence, F is continuous in the second argument and so it is l.s.c. in the second argument. Then all hypotheses of the Theorem 8 hold and hence, there exists such that

()
This completes the proof.

If we take T(x) ≡ K for all xK in Theorem 13, we have the following corollary.

Corollary 14. Let K be a nonempty compact convex subset of X and let A : KL(X, Y) be a C-strong pseudomonotone and hemicontinuous. Then, VVI has a solution.

If we set Y = and C = [0, ) in Theorem 13, we have the following result.

Corollary 15. Let K be a nonempty compact convex subset of X and let A : KL(X, ) be a monotone and hemicontinuous in the first argument. Let T : K → 2K be a set-valued mapping such that for any xK, T(x) is a nonempty convex subset of K and for any yK, T−1(y) is open in K. Assume the set P : = {xXxT(x)} is open in K and for any xK, T(x)∩{yK∣〈A(y), x − y〉>0} ≠ . Then F(T)∩VI (K, A) ≠ .

Remark 16. (1) Theorems 8 and 13 are the extensions of vector quasi-equilibrium problems and vector quasi-variational inequalities, respectively.

(2) If X is a real Banach space, then Corollary 10 comes to be Theorem  2.3 in [28].

Acknowledgment

The first author would like to thank the Office of the Higher Education Commission, Thailand for supporting by a grant fund under the program Strategic Scholarships for the Join Ph.D. Program Thai Doctoral degree for this research under Grant CHE-Ph.D.-SW-RG/41/2550, Thailand.

      The full text of this article hosted at iucr.org is unavailable due to technical difficulties.