STRONG DUALITY WITH STRICT EFFICIENCY IN VECTOR OPTIMIZATION INVOLVING NONCONVEX SET-VALUED MAPS

YU Guo-lin,ZHANG Yan,LIU San-yang

(1.Institute of Applied Mathematics,Beifang University of Nationalities,Yinchuan 750021,China)

(2.Department of Mathematics,Xidian University,Xi’an 710071,China)

余国林1,张 燕1,刘三阳2

(1.北方民族大学应用数学研究所, 宁夏 银川 750021)

(2.西安电子科技大学数学系, 陕西 西安 710071)

STRONG DUALITY WITH STRICT EFFICIENCY IN VECTOR OPTIMIZATION INVOLVING NONCONVEX SET-VALUED MAPS

YU Guo-lin1,ZHANG Yan1,LIU San-yang2

(1.Institute of Applied Mathematics,Beifang University of Nationalities,Yinchuan 750021,China)

(2.Department of Mathematics,Xidian University,Xi’an 710071,China)

This paper is diverted to the study of two strong dual problems of a primal nonconvex set-valued optimization in the sense of strict effi ciency.By using the principles of Lagrange duality and Mond-Weir duality,for each dual problem,a strong duality theorem with strict effi ciency is established.The conclusions can be formulated as follows:starting from a strictly effi cient solution of the primal problem,it can be constructed a strictly effi cient solution of the dual problem such that the corresponding objective values of both problems are equal.The results generalize the strong dual theorems in which the set-valued maps are assumed to be cone-convex.

strict effi ciency;strong duality;set-valued optimization;ic-cone-convexlikeness

1 Introduction

One of the most important topics of set-valued optimization is related to proper efficiency,this is because that the range ofthe set of(weak)effi cient solutions is often too large. In order to contract the solution range,several kinds of proper effi ciency were presented.For example,Benson effi ciency[1],Henig effi ciency[2],Geoffrion effi ciency[3],Super effi ciency [4]and Strictly effi ciency[5]etc.Especially,super effi ciency,given by Borwein and Zhuang [4],was shown to have some desirable properties.However,the condition to guarantee its existence is rather strong.Later,weakening the existence condition,Professor Cheng and Fu [5]improved the concept ofsupper effi ciency and introduced the concept of strict effi ciency.

Since duality assertions allow to study a minimization problem through a maximization problem and to know what one can expect in the bestcase.At the same time,duality resulted in many applications within optimization,and it provided many unifying conceptualinsights into economics and management science.So it is not surprising that duality is one of the important topics in set-valued optimization.There were many papers dedicated to dualitytheory ofset-valued optimization(see[6–11]).Among results obtained in this field,we want to mention the strong duality.In vector optimization,it is often said that strong duality holds between primaland dualproblems,if a weakly effi cient solution ofa primalproblem is a weakly effi cient solution ofdualproblem and such that the corresponding objective values of the primaland dualproblems are equal.If in this problem “weakly effi cient solution”is replaced by “properly effi cient solution”,then it is said that strong duality with proper effi ciency holds between the primaland dualproblems.However,strong duality with proper effi ciency was considered only for the case when proper effi ciency was understood in the sense of Geoffi ron[10]and Benson[11].

On the other hand,it is wellknown that the concept of cone-convexity and its generalizations play an important role in establishing duality theorems for set-valued optimization problems.Up to now,there are many notions of generalized convexity for set-valued maps which are introduced and are proved to be usefulfor optimization theory and related topics. Among them,the notion of ic-cone-convexlikeness seemed to be more general one[12],and was successfully applied to strict effi ciency and Henig effi ciency in set-valued optimization [13–16].

Based upon the above observation,the aim of this note is to establish the strong duality theorems with strict effi ciency for set-valued optimization problems under the ic-coneconvexlikenessassumptions.Thispaperisarranged as follows:In Section 2,some well-known definitions and results used in the sequelare recalled.In Section 3,two improved dualmodels are introduced,and strong duality theorems with strict effi ciency are established under the assumption of ic-cone-convexlikeness,respectively.

2 Preliminaries

In this paper,let X,Y and Z be real topological spaces.Let D ⊂ Y and E ⊂ Z be pointed convex cones,and denoted

Defi nition 2.1Let M be a nonempty subset of Y, ¯y ∈ M is called a minimize (maximize)point of M,if

The set of minimize(maximize)point of M is denoted by Min[M,D](Max[M,D]).

For a set A ⊂ Y,we write cone(A)={λ ·a: λ ≥ 0,a ∈ A}.The closure and interior of set A is denoted by cl(A)and int(A).A convex subset B of a cone D is a base of D if 0Y/∈ cl(B)and D=cone(B).

Throughout this paper,it is always assumed that the pointed convex cone D ⊂ Y has a base B.

Defi nition 2.2[5,13]Let M be a nonempty subset of Y, ¯y ∈ M is called a strictly minimize point of M with respect to B,if there is a neighbourhood U of 0Ysuch that

The set of strictly effi cient point of M with respect to B is denoted by Strmin[M,B].

Remark 2.1[5,13](1)With respect to the defi nition of strictly minimize points, equality(2.1)is equivalent to

Moreover,if necessary,the neighbourhood U of 0Ycan be chosen to be open,convex or balanced.

(2)Strmin[M,B]⊂ Min[M,D].

(3)Similarly, ¯y ∈ M is called a strictly maximize point of M with respect to B,ifthere is a neighbourhood V of 0Ysuch that

Remark 2.2In Defi nition 2.2,if equality(2.1)holds,then

In fact,if not,there exist λ ＞ 0,m ∈ M,d ∈ D{0Y},u ∈ U and b ∈ B,such that λ(m − ¯y+d)=u − b.Since B is the base of D,there exist µ ＞ 0 and b1∈ B such that d= µ ·b1.Since B is convex set,we get that

Therefore,we can get

which contradicts equality(2.1).

Defi nition 2.3[12]The set-valued map F:X → 2Yis called ic-D-convexlike if int(cone(im(F)+D))is convex and

where im(F)is the image of F,and that is

Assume that F:X → 2Yand G:X → 2Zare set-valued maps.This note considers the following set-valued optimization problem(SOP):

The set of feasible solution of(SOP)is denoted by Ω,that is

Defi nition 2.4If ¯x ∈ S and ¯y ∈ F(¯x) ∩ Strmin£F(S),B,then we say that(¯x,¯y)is a strictly effi cient solution of problem(SOP).

Let L(X,Y)be the family of(single-valued)linear continuous maps from X into Y.Let

Defi nition 2.5[13]Let F:X → 2Ybe a set-valued map, ¯x ∈ X and ¯y ∈ F(¯x).A map T ∈ L(X,Y)is said to be a strict subgradient of F at(¯x,¯y)if

The set of allstrict subgradients of F at(¯x,¯y)is denoted by ∂strF(¯x,¯y).

Assumption(A)[12]In problem(SOP),let ¯x ∈ S, ¯y ∈ F(¯x)and ¯z ∈ G(¯x) ∩ (−E). It is said that Assumption(A)is satisfied if there exists β ∈ [0,1)such that the set-valued map Hβ:=(F − ¯x)× (G − β ·¯z):X → 2Y×Zis ic-D × E-convexlike.

Defi nition 2.6[12]It is said that condition(CQ)holds if cl£cone(im G+E) =Z.

Lemma 2.7[13]Let ¯x ∈ S, ¯y ∈ F(¯x)and ¯z ∈ G(¯x) ∩ (−E).Let Assumption(A)and condition(CQ)be satisfied.If(¯x,¯y)is a strictly effi cient solution of problem(SOP),then there exists ¯T ∈ L+(Z,Y)such that ¯T(¯z)=0Yand

3 Strong Duality

3.1 Lagrange-Wolfe Strong Duality

We firstrewrite the Lagrange dualproblem in the form similar to the Wolfe dualproblem [17],which is denoted by problem(LWD)as follows:

Denote by Q1the set of allfeasible points of(LWD),i.e.,the set of points(ξ,u,v,T) ∈X × Y × Z × L(Z,Y)satisfying(3.1)–(3.3).Let S1be the set of all points u+T(v)such that there exists ξ∈ X with(ξ,u,v,T) ∈ Q1.

Defi nition 3.1If(ξ,u,v,T) ∈ Q1,and u+T(v) ∈ Strmax£S,B,then we say that (ξ,u,v,T)is a strictly effi cient solution of problem(LWD).

Theorem 3.2(Weak Duality)If x ∈ Ω and(ξ,u,v,T) ∈ Q1,then

ProofSince x ∈ Ω,it holds that G(x) ∩ (−E)/= ∅.So we can take a point.Hence

On the other hand,(3.2)shows that there exists a neighbourhood U of 0Ysuch that

It follows from Remark 2.2 that

So we get(3.4),as desired.

Remark 3.1In weak duality Theorem 3.2,it follows from(3.4)and Remark 2.1 that u+T(v) ∈ min£F(x),D.This leads to

so(3.4)means that y/≤ u+T(v), ∀y ∈ F(x),which is the sense of generalweak duality in literatures[6–8].

Theorem 3.3(Strong Duality)Let ¯x ∈ X, ¯y ∈ F(¯x)and ¯z ∈ G(¯x) ∩ (−E).Let Assumption(A)and condition(CQ)be satisfied.If(¯x,¯y)is a strictly effi cient solution of problem(SOP),then there exists ¯T ∈ L+(Z,Y)such that ¯T(¯z)=0,(¯x,¯y,¯z, ¯T)is a strictly effi cient solution of(LWD),and the corresponding objective values of(SOP)and(LWD)are equal.

ProofIt yields from Lemma 2.7 that there exists ¯T ∈ L+(Z,Y)such that ¯T(¯z)=0 and (¯x,¯y,¯z, ¯T) ∈ Q1.It remains to prove that ¯y= ¯y+ ¯T(¯z) ∈ Strmax[S1,B].In fact,otherwise there exist the neighbourhood U0of 0Ysuch that

Hence,there exist b0∈ (B − U0), λ ＞ 0 and ˆu+T(ˆv) ∈ S1such that b0= λ(ˆu+T(ˆv) − ¯y) or,equivalently,

This indicates that

a contradiction to the weak duality property(3.4)with x= ¯x.

3.2 Mond-Weir Strong Duality

This subsection is devoted to construct another duality problem on the basis ofthe idea of Mond-Weir[18],called the Mond-Weir duality problem(MWD),and establish a strong duality result between(SOP)and(MWD).The next problem is named the Mond-Weir dual problem of(SOP)and is denoted by(MWD):

Denote by Q2the set ofallfeasible points of(MWD),i.e.,the set ofpoints(ξ,u,v,T) ∈X × Y × Z × L(Z,Y)satisfying(3.5)–(3.8).Let S2be the set ofallpoints u such that there exists(ξ,v,T) ∈ X × Z × L(Z,Y)with(ξ,u,v,T) ∈ Q2.

Lemma 3.4It holds that Q2⊂ Q1and S2⊂ S1− D.

ProofAccording to the definitions of Q1and Q2,it is obviously that Q2⊂ Q1is satisfied.So it is to prove the second one only.Let u ∈ S2.Then there exists(ξ,v,T) ∈X × Z × L(Z,Y)such that(ξ,u,v,T) ∈ Q2⊂ Q1is satisfied.We get that

Thus,u ∈ S1− D.This completes proof.

Theorem 3.5(Weak Duality)If x ∈ Ω and(ξ,u,v,T) ∈ Q2,then there exists a neighbourhood U of 0Ysuch that

ProofBy Lemma 3.4,we obtain that Q2⊂ Q1.Again,we get from Theorem 3.2 that there exists a neighbourhood U of 0Ysuch that

Hence it follows from Remark 2.2 that

On the other hand,it yields from(3.8)that

Combing above inquality with(3.10)yields(3.9),as required.

In order to formulating the strong duality between(SOP)and(MWD),we need propose the following Lemma 3.6.

Lemma 3.6If(¯ξ,¯u,¯v, ¯T)is a strictly effi cient solution of(LWD)and ¯T(¯v)=0,then ( ¯ξ,¯u,¯v, ¯T)is a strictly effi cient solution of(MWD)and the corresponding objective values of both problems are equal.

ProofBecause(¯ξ,¯u,¯v, ¯T)is a strictly effi cient solution of(LWD),it follows from the definition of set S1that there exists a neighbourhood U of 0Ysuch that

Therefore,we get from Remark 2.2 that

On the other hand,according to Lemma 3.4,we have S2⊂ S1− D.Then we derive from ¯T(¯v)=0 that

Together(3.11)with(3.12),it is clear thatwhich is the desired result.

Theorem 3.7(Strong Duality)Let ¯x ∈ X, ¯y ∈ F(¯x)and ¯z ∈ G(¯x) ∩ (−E).Let Assumption(A)and condition(CQ)be satisfied.If(¯x,¯y)is a strictly effi cient solution of problem(SOP),then there exists ¯T ∈ L+(Z,Y)such that ¯T(¯z)=0,(¯x,¯y,¯z, ¯T)is a strictly effi cient of(MWD),and the corresponding objective values of(SOP)and(MWD)are equal.

ProofIt follows from Lemma 2.7 that there exists ¯T ∈ L+(Z,Y)such that ¯T(¯z)=0 and(¯x,¯y,¯z, ¯T) ∈ Q2⊂ Q1.Hence,we get from the strong duality Theorem 3.3 between (SOP)and(LWD)that(¯x,¯y,¯z, ¯T)is a strictly effi cient solution of(LWD)and the corresponding objective values of(SOP)and(LWD)are equal.Therefore,it yields from Lemma 3.6 that(¯x,¯y,¯z, ¯T)is also a strictly effi cient of(MWD)and the corresponding objective values of(LWD)and(MWD)are equal.This can obtain the desired results.

[1]Benson H P.An improved defi nition of proper effi ciency for vector maximization with respect to cones[J].J.Math.Anal.Appl.,1979,71:232–241.

[2]Henig M I.Proper effi ciency with respect to cones[J].J.Optim.The.Appl.,1982,36:387–407.

[3]Geoff rion A M.Proper effi ciency and the theory of vector maximization[J].J.Math.Anal.Appl., 1968,22:618–630.

[4]Borwein J M,Zhuang D M.Super effi ciency in convex vector optimization[J].Math.Meth.Oper. Res.,1991,35:175–184.

[5]Cheng Y H,Fu W T.Strong effi ciency in a locally convex space[J].Math.Meth.Oper.Res.,1999, 50:373–384.

[6]Bhatia Davinder.Lagrangian duality for preinvex set-valued functions[J].J.Math.Anal.Appl., 1997,214:599–612.

[7]Song Wen.Duality for vector optimization of set-valued functions[J].J.Math.Anal.Appl.,1996, 201:212–225.

[8]Azimov A Y.Duality for set-valued multiobjective optimization problems,part 1:mathematical programming[J].J.Optim.The.Appl.,2008,137:61–74.

[9]Frank Heydea,Carola Schrage.Continuity concepts for set-valued functions and a fundamental duality formula for set-valued optimization[J].J.Math.Anal.Appl.,2013,397:772–784.

[10]Sach P H,Lee G M,Kim D S.Strong duality for proper effi ciency in vector optimization[J].J. Optim.The.Appl.,2006,130:139–151.

[11]Sach Pham Huu,Tuan Le Anh.Strong duality with proper effi ciency in multiobjective optimization involving nonconvex set-valued maps[J].Numer.Funct.Anal.Optim.,2009,30:371–392.

[12]Sach P H.New generalized convexity notion for set-valued maps and application to vector optimization[J].J.Optim.The.Appl.,2005,125:157–179.

[13]Li Taiyong,Xu Yihong.The stictly effi cient subgradient of set-valued opttimization[J].Bull.Austr. Math.Soc.,2007,75:361–371.

[14]Yu Guolin,Liu Sanyang.Globally proper saddle point in ic-cone-convexlike set-valued optimization problems[J].Acta Math.Sinica,Eng.Ser.,2009,25:1921–1928.

[15]Yu Guolin,Lu Yangyang,Moreau-Rockafellar theorems for globally proper effi cient subgradients of set-valued maps[J].J.Math.,2012,32(6):1069–1074.

[16]Yu Guolin,Liu Sanyang.Strong effi cient solutions of nearly conesubconvexlike set-valued vector optimizaiton problems characterized by generalized saddle pointr[J].J.Math.,2008,28(1):61–66.

[17]Wolfe P.A duality theorem for nonlinear programing[J].Quart.Appl.Math.,1969,19:239–244.

[18]Weir T,Mond B.Pre-invex functions in multiple objective optimization[J].J.Math.Anal.Appl., 1988,136:29–38.

非凸集值优化问题严有效解的强对偶定理

本文研究了非凸集值向量优化的严有效解在两种对偶模型的强对偶问题.利用Lagrange对偶和Mond-Weir对偶原理, 获得了如下结果: 原集值优化问题的严有效解, 在一些条件下是对偶问题的强有效解,并且原问题和对偶问题的目标函数值相等;推广了集值优化对偶理论在锥-凸假设下的相应结果.

严有效性;强对偶;集值优化;生成锥内部凸-锥类凸性

类 号:90C29;90C46

O224

余国林1,张 燕1,刘三阳2

(1.北方民族大学应用数学研究所, 宁夏 银川 750021)

(2.西安电子科技大学数学系, 陕西 西安 710071)

tion:90C29;90C46

A < class="emphasis_bold">Article ID:0255-7797(2017)02-0223-08

0255-7797(2017)02-0223-08

∗Received date:2015-01-27 Accepted date:2015-09-24

Foundation item:Supported by Natural Science Foundation of China(11361001);Natual Science Foundation of Ningxia(NZ14101).

Biography:Yu Guolin(1974–),male,born at Yinchuan,Ningxia,professor,major in optimization theory and applications,nonlinear analysis.