• Aucun résultat trouvé

WEIGHTED VARIATIONAL INEQUALITIES WITH SET-VALUED MAPPINGS

N/A
N/A
Protected

Academic year: 2022

Partager "WEIGHTED VARIATIONAL INEQUALITIES WITH SET-VALUED MAPPINGS"

Copied!
13
0
0

Texte intégral

(1)

WEIGHTED VARIATIONAL INEQUALITIES WITH SET-VALUED MAPPINGS

MIRUNA BELDIMAN

We consider some systems of vector general variational inequalities and weighted variational inequalities, and find equivalence conditions between them. Then we establish for these classes of problems a few existence results under different types of generalized weighted monotonicity assumptions.

AMS 2000 Subject Classification: 47J20, 58E35.

Key words: weighted general variational inequalities, system of vector variational inequalities, generalized weighted pseudomonotonicity, hemicontinu- ity, convexity.

1. INTRODUCTION

Because of their applications in economics, game theory, mathematical physics, operations research and other areas, many classes of vector variational inequalities were intensively studied. For existence of solutions, resolution methods or equivalence with equilibrium and optimization problems see, for example, [11], [14], [15], [19], [17] and the references therein. For the study of variational inequalities systems in the single valued case, see [16] and [18].

Ansari, Schaible and Yao [4] introduced a system of vector equilibrium problems with vector valued bifunction defined over a product set, which in- cludes as particular cases systems of generalized vector variational inequalities and optimization problems.

The same authors [5] defined another systems of vector equilibrium prob- lems and obtained an existence theorem based on a known maximal element theorem. From this, they derived existence results for some systems of vector variational-like inequalities and systems of implicit vector variational inequal- ities.

Ansari and Khan [2] proved the existence of solution for a variational inequality over product sets, which is equivalent to a system of vector vari- ational inequalities, under generalized pseudo-monotonicity assumptions, in the meaning of Brezis [8]. They used a fixed point theorem of Chowdury and Tan [9].

REV. ROUMAINE MATH. PURES APPL.,52(2007),3, 315–327

(2)

Later, Allevi, Gnudi and Konnov [1] obtained existence results for a system of general vector variational inequalities and countable product of sets with set-valued mappings in real Banach spaces, using the classical Ky Fan lemma [13].

Ansari, Khan and Siddiqi [3] introduced weighted variational inequalities which involve set- valued mappings and established the relationship between systems of general vector variational inequalities and systems of weighted vari- ational inequalities, which can be written as duality products.

Following this direction, we extend their results for arbitrary spaces and general mappings between them. We define some systems of general vector variational inequalities and weighted variational inequalities and give equiva- lence conditions. Also, we obtain existence of solutions using a theorem proved in [7] for the single- valued case.

2. DEFINITIONS AND SOME PRELIMINARY RESULTS For each given m∈N, we denote byRm+ the nonnegative orthant ofRm, that is,

Rm+ ={u= (u1, . . . , um)Rm :uj 0, forj= 1, . . . , m} and let

T+m=

u= (u1, . . . , um)Rm+ : m j=1

uj = 1

be the simplex of Rm+ . In what follows we will shorten “with respect to”

as “wrt”.

LetI ={1, . . . , n}be a finite index set. For eachi∈I letibe a positive integer, Xi a real topological vector space (not necessarily Hausdorff), Ki a nonempty convex subset ofXi, and Yi an arbitrary set. Put X =

iIXi and K=

iIKi.

For each x X, xi Xi stands for ith coordinate of x and we write x= (xi)iI. For eachi∈I letfi :K →Yi and Ψi:Yi×Ki×Ki Ri be two maps and Ψ = (Ψi)iI. Assume that W = (W1, . . . , Wn) n

i=1

R+i\ {0}

is a given weight vector and consider the following problems, where “·” stands for the inner product onR+i.

(Ψ-WVIP) Findx∈K such that

iI

Wi·Ψi(fi(x), xi;yi)0 for allyi ∈Ki,i∈I.

(3)

(Ψ-SWVI) Findx∈K such that,

Wi·Ψi(fi(x), xi;yi)0 for allyi ∈Ki,i∈I.

We denote by Kw (respectively Ksw) the solution set of (Ψ-WVIP) (respectively, (Ψ-SWVI)).

Definition 2.1 ([7]). A family (fi)iI of functions is said to be (i) weighted monotone wrt(W,Ψ) if

iI

Wi·i(fi(y), xi;yi)Ψi(fi(x), xi;yi))0

for allx, y∈K, andweighted strictly monotone wrt(W, ψ) if the inequality is strict for allx=y;

(ii)weighted pseudomonotone wrt (W,Ψ) if

iI

Wi·Ψi(fi(x), xi;yi)0

iI

Wi·Ψi(fi(y), xi;yi)0

for allx, y∈K, andweighted strictly pseudomonotone wrt(W,Ψ) if the second inequality is strict for all x=y;

(iii)weighted maximal pseudomonotone wrt(W,Ψ) if it is weighted pseu- domonotone wrt (W,Ψ) and

iI

Wi·Ψi(fi(z), xi;zi)0, ∀z∈(x, y]

iI

Wi·Ψi(fi(x), xi;yi)0 for allx, y∈K, where (x, y] =

iI(xi, yi], andweighted maximal strictly pseu- domonotone wrt(W,Ψ) if it is weighted strictly pseudomonotone wrt (W,Ψ) and (3.1) holds.

Definition 2.2 ([7]). A family (fi)iI of functions is said to be weighted hemicontinuous wrt (W,Ψ) if for all x, y K the mapping λ [0,1]

iI

Wi·Ψi(fi(x+λ(y−x)), xi;yi) is continuous.

Definition 2.3 ([7]). A family (fi)iI is said to be weighted B-pseudo- monotone wrt(W,Ψ) if for eachx∈Kand every net{xα}α∈ΓinKconverging to xwith

lim sup

α iI

Wi·Ψi(fi(xα), xαi;xi) 0 we have

lim sup

α iI

Wi·Ψi(fi(xα), xαi;yi)

iI

Wi·Ψi(fi(x), xi;yi), for ally∈K.

(4)

In what follows we shall use Theorem2.1 ([7]). Assume that

(i1)the family (fi)iI of functions is weighted maximal pseudomonotone wrt (W,Ψ);

(i2) there exists a nonempty closed compact subset D of K and y∈D such that

iI

Wi·Ψi(fi(x), xi;yi)>0 for all x∈K\D;

(i3) the mapping y→

i Wi·Ψi(fi(x), xi;yi) is convex on K;

(i4)

iIWi·Ψi(fi(x), xi;xi) = 0 for all x∈K;

(i5)for x, y∈coAand every sequence {xα}α∈Γin K converging to xwe have lim

α∈Γ

iIWi·Ψi(fi(xα), xαi;yi) =

iIWi·Ψi(fi(x), xi;yi).

Then there exists a solution x K of (Ψ-WVIP), hence a solution of (Ψ-SWVI). Furthermore, if W

iIT+i, then there exists a normalized solution x K of (Ψ-WVIP), hence a solution of (Ψ-SVVI)w. Finally, if W n

i=1

intT+i

,then x ∈K is a solution of (Ψ-SVVI).

Corollary2.1 [7]. Assume that conditions (i2)–(i5) from Theorem2.1 and (i6) below hold:

(i6) the family {fi}iI of functions is weighted hemicontinuous and weighted pseudomonotone wrt (W,Ψ).

Then there exists a solution x K of (Ψ-WVIP),hence a solution of (Ψ-SWVI). If W

iIT+i then there exists a normalized solution x∈K of (Ψ-WVIP) hence a solution of (Ψ-SVVI)w. Moreover, if W n

i=1

intT+i then x∈K is a solution of(Ψ-WVIP),hence a solution of (Ψ-SVVI).

3. A CLASS OF WEIGHTED VARIATIONAL INEQUALITIES WHICH INVOLVE SET-VALUED MAPPINGS

In this section we introduce a class of generalized weighted variational inequalities over products of sets. It includes as particular cases the weighted variational inequalities for multivalued maps and systems of weighted varia- tional inequalities from [3] as well as the weighted Ψ-variational inequalities and Ψ-systems of vector variational inequalities from [7].

(5)

A few extensions of weighted monotonicity concepts are also considered, and under such assumptions some existence results for our generalized weighted variational problem and systems of generalized vector variational inequalities are obtained.

Fori∈I and j = 1,2, . . . , i,i N, letFij :K 2Yi be a multivalued map with nonempty values. Fori∈I set

Fi(x) =Fi1(x)× · · · ×Fii(x), x∈K.

Also, put

F(x) ={Fi(x)}iI.

Consider the following general systems of generalized vector variational inequalities.

(Ψ-SGVVI) Find x∈K such that

∃ui∈ Fi(x) with Ψi(ui, xi;yi)∈/R+i\ {0}, ∀yi∈Ki, i∈I, and

(Ψ-SGVVI)w Find x∈K such that

∃ui ∈ Fi(x) with Ψi(ui, xi;yi)∈/R+i, ∀yi∈Ki i∈I.

Also, consider the system:

(Ψ-SGVVI) Find x∈K such that

∃ui∈ Fi(x) with Ψi(ui, xi;yi)∈/R+i, ∀yi∈Ki, i∈I.

Remark 3.1. Every solution of (Ψ-SGVVI) is a solution of (Ψ-SGVVI)w and every solution of (Ψ-SGVVI)w is a solution of (Ψ-SGVVI),but the con- verse implications are not true in general.

Thus (Ψ-SGVVI) is a weak formulation of (Ψ-SGVVI)w. Remark 3.2. If for everyi∈I, Yi =Xi and

Ψi(ui, xi;yi) =

ui1, xi−yi, . . . ,uii, xi−yi ,

where ·,· denotes the continuous pairing between Xi and Xi, then (Ψ-SGVVI),(Ψ-SGVVI)wand (Ψ-SGVVI)lead to (SGVVI),(SGVVI)wfrom [3] and to the problem of systems of generalized vector variational inequalities introduced and studied by Ansari, Schaible and Yao [5],respectively.

Now, to solve (Ψ-SGVVI) and (Ψ-SGVVI)w, we consider as in [3] a weighted variational inequality problem type. So, we introduce the following weighted generalized variational inequality problem over products of sets:

(Ψ-WGVIP) Find x∈K and u∈ F(x) such that

iI

Wi·Ψi(u, xi;yi)0, ∀yi ∈Ki,i∈I.

(6)

We denote the solution of (Ψ-WGVIP) byKwg.

Also consider the problem of systems of weighted generalized variational inequalities

(Ψ-SWGVI) Find x∈K and u∈ F(x) such that Wi·Ψi(ui, xi;yi)0, ∀yi∈Ki, i∈I.

Note that if Wi T+i for each i∈ I, then the solution of (Ψ-WGVIP) or (Ψ-SWGVI) is said to be normalized. Denote by Kswg the solution set of (Ψ-SWGVI) and byKsnwg the normalized solution set of (Ψ-SWGVI).

The following definitions can be seen as an extension of relative B-pseudo- monotonicity introduced in [2], weighted B-pseudomonotonicity for a family of functions given in [3] and of Definition 2.3.

Definition 3.1. A family{Fij} iI

j=1,i is said to be

(i) weighted generalized pseudomonotone wrt (W,Ψ) if for all x, y K andu∈ F(x), v∈ F(y) we have

iI

Wi·Ψi(ui, yi;xi)0

iI

Wi·Ψi(vi, yi;xi)0;

(ii) weighted generalized maximal pseudomonotone wrt (W,Ψ) if it is weighted generalized pseudomonotone wrt (W,Ψ) and for all x, y K and u∈ F(x) there exists δ0(0,1] such that

iI

Wi·Ψi(si, zi;xi)0,∀si∈ Fi(z) ,i∈I,z∈(x, y]

iI

Wi·Ψi(ui, yi;xi)0, wherez=x+t(y−x),0< t≤δ0;

(iii)weightedu-hemicontinuous wrt(W,Ψ) if for allx, y∈Kthe mapping λ∈[0,1]

iIWi·Ψi(si(x+λ(y−x)), yi;xi) is upper semicontinuous at 0, wheresi(x+λ(y−x))∈ Fi(x+λ(y−x)), i∈I.

Remark 3.3. In the case considered in Remark 3.2,we obtain the defini- tions of generalized weighted pseudomonotonicity wrtW,generalized weighted maximal pseudomonotonicity wrt W and weighted u-hemicontinuity respec- tively (see [3]).

Proposition 3.1. Assume that

•the family {Fij} iI

j=1,i is hemicontinuous wrt (W,Ψ)and weighted gen- eralized pseudomonotone wrt (W,Ψ) ;

(7)

there exist δ > 0 and τ > 0 such that for any i I, x, y K and t∈(0, δ) we have

Ψi(ui, xi+t(yi−xi) ;xi) =tτΨi(ui, yi;xi). Then the family{Fij} i∈I

j=1,i is weighted generalized maximal pseudomono- tone wrt (W,Ψ).

Proof. Letx, y∈K such that for anyu ∈ F(x) we have

iI

Wi·Ψi(ui, yi;xi)>0.

By the weighted u-hemicontinuity of the family {Fij} iI

j=1,i,there exists δ,0< δ 1,such that

iI

Wi·Ψi(si(x+t(y−x)), yi;xi)>0

for all s(x+t(y−x)) ∈ F(x+t(y−x)) and t (0, δ). This inequality together with (j2) yields

iI

Wi·Ψi(si(x+t(y−x)), xi+t(yi−xi) ;xi)>0 for any t∈(0,min{δ, δ}). The proof is complete.

Corollary 3.1 ([3]). If the family {Fij} iI

j=1,i of multivalued maps Fij : K 2Xi is u-hemicontinuous and weighted generalized pseudomono- tone wrt W, then it is weighted generalized maximal pseudomonotone wrt the same weight vector W.

In the multivalued case we need some results about continuous selections.

Definition 3.2 ([20]). Consider two topological vector spaces S and Z, a subsetU of S, a multivalued map G:U 2Z\ {Φ} and g :U Z. We say thatg is a selection of Gon U if g(x) ∈G(x) for allu ∈U. The selection g is said to be a continuous selection of G on U if it is continuous on U and a selection ofGon U.

For some results on the existence of continuous selections see [10,20]. It is easy to prove the following results.

Lemma3.1. Assume that

•for each i∈I and j= 1,2, . . . , i, fij is a selection of Fij on K;

•the family {Fij} iI

j=1,i is weighted generalized maximal pseudomonotone wrt (W,Ψ).

(8)

Then the family {fij} iI

j=1,i

is weighted maximal pseudomonotone wrt (W, Ψ).

Lemma 3.2. Assume that fij is a selection of Fij on K for each i∈ I and j = 1,2, . . . , i. Let x K and let u be given by uij Fij(x) for each i∈I and j= 1,2, . . . , i. Then

•x∈Kw (x, u)∈Kwg;

•x∈Ksw (x, u)∈Kswg.

Lemma3.3. Under the assumptions of Lemma 3.1,

•if W n

i=1T+i and x∈Ksw,then (x, u) is a solution of (Ψ-SGVVI)w;

•if W n

i=1

intT+i

andx∈Ksw,then(x, u)is a solution of (Ψ-SGVVI).

As for the existence of a solution of (Ψ-SGVVI),we have the following results.

Theorem3.1. Assume that (j1) the family {Fij} iI

j=1,i is weighted generalized maximal pseudomono- tone wrt (W,Ψ) ;

(j2) for each i∈I, j = 1, . . . , i there exists a selection (not necessarily continuous)fijof Fij onKfor which assumptions (i1)(i3)from Theorem2.1 are satisfied;

(j3) there exists a nonempty closed compact subset D of K and y∈D such that

iI

Wi·Ψi(ui, xi;yi)>0

for all ui∈ Fi(x), i∈I,andx∈K\D.

Under these assumptions (k1) Kwg=∅ and Kswg=∅;

(k2) if W

iIT+i then the solution set of (Ψ-SGVVI)w is nonempty;

(k3)if W

iI

intT+i

then the solution set of (Ψ-SGVVI)is nonempty.

Proof. By (j2),for i∈ I and j = 1, . . . , i there is a function fij such that fij(x) Fij(x) for all x K. Using Lemma 3.1 we get that

fij is weighted maximal pseudomonotone wrt (W,Ψ). Now, we see that the as-i,j

sumptions of Theorem 2.1 are satisfied for{fi}i,so that Kw =Ksw =. Let x Ksw. Put uij =fij(x) for i ∈I and j = 1, . . . , i. Hence ui = fi(x) Fi(x) and Lemma 3.2 implies that (x, u) Kwg and (x, u) Kswg, whereu= (ui)iI.

(9)

For (k2) and (k3) we see that (x, u) is a normalized solution of (Ψ-SWGVI).

Now,using Lemma 3.3 we obtain that (x, u) is a solution of (Ψ-SGVVI)w,in case (k2), and a solution of (Ψ-SGVVIP) in case (k3). Thus, the proof is complete.

Using Lemma 3.1 we have

Corollary 3.2. Assume that (j2) and (j3) from Theorem 3.1 hold as well as

(j4) the family {Fij} iI

j=1,i is u-hemicontinuous and weighted generalized pseudomonotone wrt (W,Ψ).

Then the conclusions of Theorem 3.1 hold.

Corollary3.3 ([3]). Let the family Fij

iI

j=1,i of multivalued maps be weighted generalized maximal pseudomonotone wrt W. Asume that

for each i I and each j = 1, . . . , i there exists a selection (not necessarily continuous) fij of Fij on K;

•there exists a nonempty closed compact subset Dof Kand y∈Dsuch that

iIWi· ui, xi−yi>0 for allx∈K\D,andui ∈ Fi(x),i∈I.

Then the solution set Kwg of (WGVIP) is nonempty and so is Kswg. Furthermore, if W

iIT+i (respectively W

iI

intT+i

) then the solution set of (SGVVI)w (respectively (SGVVI))is nonempty.

Theorem3.2. Assume (i3)from Theorem 2.1 and that (j5) the family

Fij

iI

j=1,i is weighted generalized pseudomonotone wrt (W,Ψ) ;

(j6) for each i I and j = 1, . . . , i there exists a continuous selection fij on K for which assumptions (i3)(i5) from Theorem 2.1 are satisfied.

Then the conclusions of Theorem 3.1 hold.

Proof. According to (j6) we have a family of continuous functions fij

i,j

such that fij(x) Fij(x) for all x K. By (i4) and Lemma 3.1, the fami- ly

fij

i,j of continuous functions is weighted pseudomonotone wrt (W,Ψ).

Hence this family is weighted pseudomonotone and weighted hemicontinuous wrt (W,Ψ). Now, we can use Corollary 2.1. Therefore, there existsx ∈Kw, hencex∈Ksw.

If we put uij = fij(x) Fij(x) for i I, j = 1, . . . , i, we can use Lemma 3.2 and then get that (x, u) Kwg and (x, u) Kswg, where u= (ui)iI.

For the last part of this theorem we use Lemma 3.3. The proof is com- plete.

(10)

Now, by Proposition 3.1 and Lemma 3.1 we obtain

Corollary 3.4. Assume (i1) from Corollary 3.2 and (i2), (i3) from Theorem3.2 Then (k1)(k3) from Theorem 3.2 hold.

Corollary3.5 ([3]).Let the family Fij

iI

j=1,i of multivalued maps be weighted generalized pseudomonotone wrt W. Assume that

•for each i∈I and each j= 1, . . . , i there exists a continuous selection fij of Fij on K;

•there exists a nonempty closed compact subset Dof Kand y∈Dsuch that

iIWi· ui, xi−yi>0,for all x∈K\D and ui ∈ Fi(x), i∈I. Then the solution set Kwg of (WGVIP) is nonempty and so is Kswg. Furthermore,if W

iIT+i(respectively,W

iI

intT+i

),then the solution set of (SGVVI)w (respectively, (SGVVI)) is nonempty.

Corollary3.6 ([3]).Let the family Fij

iI

j=1,i of multivalued maps be weightedu-hemicontinuous and weighted generalized pseudomonotone wrt W. Assume that

for each i I and each j = 1, . . . , i there exists a selection (not necessarily continuous) fij of Fij on K;

•there exists a nonempty closed compact subset Dof K and y∈Dsuch that

iIWi· ui, xi−yi>0 for all x∈K\D andui ∈ Fi(x),i∈I.

Then the solution set Kwg of (WGVIP) is nonempty and so is Kswg. Furthermore, if W

iIT+i (respectively W

iI

intT+i

) then the solution set of (SGVVI)w (respectively (SGVVI)) is nonempty.

Corollary3.7 ([3]). Let the family Fij

iI

j=1,i of multivalued maps be weighted generalized pseudomonotone wrt W. Assume that

•for each i∈I and each j= 1, . . . , i there exists a continuous selection fij of Fij on K;

•there exists a nonempty closed compact subset Dof K and y∈Dsuch that

iIWi· ui, xi−yi>0 for all ui ∈ Fi(x),i∈I,x∈K\D.

Then, the solution set Kwg of (WGVIP) is nonempty and so is Kswg. Furthermore,if W

iIT+i then the normalized solution set of (SGVVI)w is nonempty.

To conclude, we consider the case of multivalued maps of B-pseudo- monotone type.

(11)

Definition 3.3. A family Fij

iI

j=1,i is said to beweighted generalized B- pseudomonotone wrt (W,Ψ) if for each x K and every net {xα}α∈Γ in K converging tox with

lim inf

α∈Γ

iI

Wi·Ψi(uαi, xαi;xi)0, ∀uαi ∈Fi(xα)

we have

iI

Wi·Ψi(ui, xi;yi)lim sup

α∈Γ

iI

Wi·Ψi(uαi, xαi;yi) ,

for allui∈Fi(x), uαi ∈Fi(xα),i∈I, andy ∈K.

Remark 3.4. If we take Ψi as in Remark 3.2 we obtain the definition of weighted generalizedB-pseudomonotonicity wrtW from [3].

We easily get

Lemma3.4. Assume that

the family Fij

iI

j=1,i is weighted generalized B-pseudomonotone wrt (W,Ψ) ;

for each i I and j = 1, . . . , i, fij : K Yi is a selection of Fij : K→2Yi on K.

Then the family fij

i,j is weighted B-pseudomonotone wrt (W,Ψ).

Now, from Lemma 3.4 and Theorem 3.2 we obtain Theorem3.3. Assume (i3) from Theorem 3.1 and that

the family Fij

iI

j=1,i is weighted generalized B-pseudomonotone wrt (W,Ψ) ;

•for each i∈I, j = 1, . . . , i and all A∈ F(K)there exists a continuous selection fij of Fij on coA which satisfies the assumptions (i3)–(i5) from Theorem 2.1.

Then (k1)–(k3) from Theorem 3.1 are true.

Corollary3.8 ([3]). Let the family Fij

iI

j=1,i of multivalued maps be weighted generalized B-pseudomonotone wrt W. Assume that

for each i I, j = 1, . . . , i and for all A ∈ F(K) there exists a continuous selection fij of Fij on coA;

•there exists a nonempty closed compact subset Dof Kand y∈Dsuch that

iIWi· ui, xi−yi>0,for all ui∈ Fi(x), i∈I,and x∈K\D.

(12)

Then the solution set Kwg of (WGVIP) is nonempty and so is Kswg. Furthermore,if W

iIT+i (respectively, W

iI

intT+i

),then the solution set of (SGVVI)w (respectively, (SGVVI)) is nonempty.

REFERENCES

[1] E. Allevi, A. Gnudi and I.V. Konnov, Generalized vector variational inequalities over countable product of sets. J. Global Optim.30(2004), 155–167.

[2] Q.H. Ansari and Z. Khan, Relatively B-pseudomonotone variational inequalities over product of sets. J. Ineq. Pure Appl. Math.4(2003), Art. 6.

[3] Q.H. Ansari, Z. Khan and A.H. Siddiqi, Weighted variational inequalities. J. Optim.

Theory Appl.127(2005), 263–283.

[4] Q.H. Ansari, S. Schaible and J.C. Yao, System of vector equilibrium problems and its applications. J. Optim. Theory Appl.107(2000), 547–557.

[5] Q.H. Ansari, S. Schaible and J.C. Yao,System of generalized vector equilibrium problems with applications. J. Global Optim.22(2002), 3–16.

[6] Q.H. Ansari and J.C. Yao, A fixed-point theorem and its applications to the system of variational inequalities. Bull. Austral. Math. Soc.59(1999), 433–442.

[7] M. Beldiman and V. Preda,On weighted variational inequalities. Submitted.

[8] H. Br´ezis,Equations et in´equations non lineaires dans les espaces vectoriels en dualit´e.

Ann. Inst. Fourier (Grenoble)18(1968), 115–175.

[9] M.S.R. Chowdury and K.K. Tan,Generalized variational inequalities for quasimonotone operators and applications. Bull. Polish Acad. Sci. Math.45(1997), 25–54.

[10] X.P. Ding, W.K. Kim and K.K. Tan, A selection theorem and its application. Bull.

Austral. Math. Soc.46(1992), 205–212.

[11] F. Giannessi, Theorems of the alternative, quadratic programs and complementarity problems. In: R.W. Cottle, F. Giannessi, J.L. Lions (Eds.) Variational Inequalities and Complementarity Problems, pp. 151–186. Wiley, New York, 1998.

[12] I.V. Konnov,Relatively monotone variational inequalities over product sets. Oper. Res.

Lett.28(2001), 21–26.

[13] K. Fan,A generalization of Tychonoff ’s fixed-point theorem. Math. Ann.142(1961), 305–310.

[14] S.K. Mishra and M.A. Noor, On vector variational-like inequality problems. J. Math.

Anal. Appl. (to appear).

[15] S.K. Mishra and S.Y. Wang,Vector variational-like inequalities and non-smooth vector optimization problems. Nonlinear Anal.64(2006), 1939–1945.

[16] A. Nagurney, Network Economics: A Variational Inequality Approach. Kluwer, Dor- drecht, 1993.

[17] M.A. Noor and K.I. Noor,On generalized mixed quasivariational inequalities. J. Optim.

Theory Appl.120(2004), 579–599.

[18] J.S. Pang,Asymmetric variational inequalities over product of sets.Math. Programming 31(1985), 206–219.

(13)

[19] V. Preda, M. Beldiman and A. Batatorescu,On variational-like inequalities with gen- eralized monotone mappings. In:Proc.8th Internat. Sympos. Generalized Convexity and Monotonicity. Springer, 2006.

[20] D. Repov and P.V. Semenov,Continuous Selection of Multivalued Mappings.Kluwer, Dordrecht, 1998.

Received 18 July 2006 University of Bucharest

Faculty of Mathematics and Computer Science Str. Academiei 14

010014 Bucharest, Romania

Références

Documents relatifs

The initial motivation which eventually led to the formulation of the above conjecture stems from the problem of stable representation of positively homoge- neous polyhedral

The remainder of the paper is devoted to show that in every normed space of dimension at least 2 one can construct a quasi-contraction having a non- connected set of fixed points..

2014;419:904–937], we establish a general equilibrium version of set- valued Ekeland variational principle (denoted by EVP), where the objective bimap is defined on the product

While existing approaches based on deformable models applied to vector- valued images exploit local structure information in a scalar way to define vector edges, the 4DGVF field

that for multifunctions d~ associated with convex functions and saddle functions, or more generally for any maximal monotone multifunction; the Clarke tangent cone to

Unité de recherche INRIA Rennes : IRISA, Campus universitaire de Beaulieu - 35042 Rennes Cedex (France) Unité de recherche INRIA Rhône-Alpes : 655, avenue de l’Europe -

Based on the projection method, in this paper we consider the approxi- mate solvability of a system of nonlinear relaxed cocoercive variational inequal- ities in a Hilbert

We introduce an iterative method for finding a common fixed point of a semigroup of nonexpansive mappings in a Hilbert space, with respect to a sequence of left regular means defined