Abstract
In this paper, a class of E-differentiable vector optimization problems with both inequality and equality constraints is considered. The so-called vector mixed E-dual problem is defined for the considered E-differentiable vector optimization problem with both inequality and equality constraints. Then, several mixed E-duality theorems are established under (generalized) V-E-invexity hypotheses.
We’re sorry, something doesn't seem to be working properly.
Please try refreshing the page. If that doesn't work, please contact support so we can address the problem.
Avoid common mistakes on your manuscript.
1 Introduction
Multiobjective programming duality has been of much interest in the recent past. Many authorious classes of multiobjective programming problems (see, for example, [1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18], and others). Youness [19] introduced the concepts of an E-convex set and an E-convex function. Megahed et al. [20] presented a new concept of an E-differentiable convex function and they established optimality conditions for mathematical programming problems in which the functions involved are E-differentiable. Abdulaleem [21] has introduced a new concept of generalized convexity as a generalization of the notion of E-differentiable E-convexity and the notion of differentiable invexity introduced by Hanson [22]. Namely, Abdulaleem defined the concept of E-differentiable E-invexity in the case of (not necessarily) differentiable vector optimization problems with E-differentiable functions. Recently, Abdulaleem [4] introduced a new concept of generalized convexity as a generalization of the E-differentiable E-invexity notion and the concept of V-invexity given by Jeyakumar and Mond [23]. Namely, Abdulaleem defined the concept of E-differentiable V-E-invexity in the case of (not necessarily) differentiable vector optimization problems with E-differentiable functions and used this concept to prove sufficient optimality conditions for a new class of nonconvex E-differentiable vector optimization problems.
In this paper, a class of E-differentiable V-E-invex vector optimization problems with both inequality and equality constraints are considered. A mixed E-dual problem is defined for the considered E-differentiable V-E-invex vector optimization problem with both inequality and equality constraints. Then, various mixed E-duality theorems are established between the considered E-differentiable multicriteria optimization problem and its vector mixed E-dual problem under appropriate (generalized) V-E-invexity hypotheses.
2 Definitions and Preliminaries
Let \(R^{n}\) be the n-dimensional Euclidean space and \(R_{+}^{n}\) be its nonnegative orthant. The following convention for equalities and inequalities will be used in the paper. For any vectors \(x=\left( x_{1},x_{2},...,x_{n}\right) ^{T}\) and \(y=\left( y_{1},y_{2},...,y_{n}\right) ^{T}\) in \(R^{n}\), \(x>y\) means that the vector x is componentwise greater than the vector y. Similarly, the same convention has been used for \(x=y\), \(x \geqq y\) and \(x \ge y.\)
First, we recall for a common reader the definition of E-differentiable function introduced by Megahed et al. [20].
Definition 1
[20] Let \(E:R^{n}\rightarrow R^{n}\) and \(f:R^{n}\rightarrow R\) be a (not necessarily) differentiable function at a given point \(u\in R^{n}\). It is said that f is an E-differentiable function at u if and only if \(f\circ E\) is a differentiable function at u (in the usual sense), that is,
where \(\theta \left( u,x-u\right) \rightarrow 0\) as \(x\rightarrow u\).
Now, we give the definition of a V-E-invex function introduced by Abdulaleem [4].
Definition 2
[4] Let \(E:R^{n}\rightarrow R^{n}\) and \(f:R^{n}\rightarrow R^{k}\) be an E-differentiable function on \(R^{n}\). It is said that f is a V-E-invex function (a strictly V-E-invex function) with respect to \(\eta\) at \(u\in R^{n}\) on \(R^{n}\) if, there exist functions \(\eta :R^{n}\times R^{n}\rightarrow R^{n}\) and \(\alpha _{i}: R^{n}\times R^{n}\rightarrow R_{+}\setminus \{0\},\) \(i=1,2,...,k,\) such that, for all \(x\in R^{n}\) \((E(x)\ne E(u))\), the inequalities
hold. If inequalities (2) are fulfilled for any \(u\in R^{n}\) \((E(x)\ne E(u))\), then f is a V-E-invex (strictly V-E-invex) function at u on \(R^{n}\). Each function \(f_{i}\), \(i = 1,...,k,\) satisfying Eq. (2), is said to be \(\alpha _{i}\)-E-invex (strictly \(\alpha _{i}\)-E-invex) with respect to \(\eta\) at u on \(R^{n}\).
Remark 1
Note that the Definition 2 generalizes and extends several generalized convexity notions, previously introduced in the literature. Indeed, there are the following special cases:
-
(a)
In the case when \(\alpha _{i}(x,u)=1\), \(i=1,...,k,\) then the definition of a V-E-invex function reduces to the definition of an E-invex function introduced by Abdulaleem [21].
-
(b)
If f is differentiable and \(E(x)\equiv x\) (E is an identity map), then the definition of a V-E-invex function reduces to the definition of a V-invex function introduced by Jeyakumar and Mond [23].
-
(c)
If f is differentiable, \(E(x)\equiv x\) (E is an identity map) and \(\alpha _{i}(x,u)=1\), \(k=1\), then the definition of a V-E-invex function reduces to the definition of an invex function introduced by Hanson [22].
-
(d)
If f is defined by \(\eta (x,u)=x-u\) and \(\alpha _{i}(x,u)=1,\) \(i = 1,...,k,\) then we obtain the definition of an E-differentiable E-convex vector-valued function introduced by Megahed et al. [20].
-
(e)
If f is differentiable, \(E(x)=x\) and \(\eta (x,u)=x-u\) and \(\alpha _{i}(x,u)=1,\) \(i = 1,...,k,\) then the definition of a V-E-invex function reduces to the definition of a differentiable convex vector-valued function.
-
(f)
If f is a differentiable scalar function, \(\eta (x,u)=x-u\) and \(\alpha _{i}(x,u)=1\), then we obtain the definition of a differentiable E-convex function introduced by Youness [19].
Now, we give various classes of generalized E-differentiable V-E-invex functions.
Definition 3
Let \(E:R^{n}\rightarrow R^{n}\) and \(f:R^{n}\rightarrow R^{k}\) be an E-differentiable function on \(R^{n}\). It is said that f is a V-E-pseudo-invex function with respect to \(\eta\) at \(u\in R^{n}\) on \(R^{n}\) if, there exist functions \(\eta :R^{n}\times R^{n}\rightarrow R^{n}\) and \(\alpha _{i}: R^{n}\times R^{n}\rightarrow R_{+}\setminus \{0\},\) \(i=1,2,...,k,\) such that, for all \(x\in R^{n},\) the relations
hold. If Eq. (3) are fulfilled for any \(u\in R^{n}\), then f is V-E-pseudo-invex with respect to \(\eta\) on \(R^{n}\). Each function \(f_{i}\), \(i = 1,...,k,\) satisfying Eq. (3) is said to be \(\alpha _{i}\)-E-pseudo-invex with respect to \(\eta\) at u on \(R^{n}\).
Definition 4
Let \(E:R^{n}\rightarrow R^{n}\) and \(f:R^{n}\rightarrow R^{k}\) be an E-differentiable function on \(R^{n}\). It is said that f is a V-E-quasi-invex function with respect to \(\eta\) at \(u\in R^{n}\) on \(R^{n}\) if there exist functions \(\eta :R^{n}\times R^{n}\rightarrow R^{n}\) and \(\alpha _{i}: R^{n}\times R^{n}\rightarrow R_{+}\setminus \{0\},\) \(i=1,2,...,k,\) such that, for all \(x\in R^{n},\) the relations
hold. If Eq. (4) are fulfilled for any \(u\in R^{n}\), then f is V-E-quasi-invex on \(R^{n}\). Each function \(f_{i}\), \(i = 1,...,k,\) satisfying Eq. (4) is said to be \(\alpha _{i}\)-E-quasi-invex with respect to \(\eta\) at u on \(R^{n}\).
In this paper, we consider the following (not necessarily differentiable) multiobjective programming problem (MOP) with both inequality and equality constraints defined as follows:
where \(f_{i}:R^{n}\rightarrow R\), \(i\in I=\left\{ 1,...,p\right\}\), \(g_{j}:R^{n}\rightarrow R\), \(j\in J\), \(h_{t}:R^{n}\rightarrow R\), \(t\in T\), are real-valued functions defined on \(R^{n}.\) We shall write \(g:=\left( g_{1},...,g_{m}\right) :R^{n}\rightarrow R^{m}\) and \(h:=\left( h_{1},...,h_{q}\right) :R^{n}\rightarrow R^{q}\) for convenience. Let
be the set of all feasible solutions of (MOP). Further, we denote by \(J\left( x\right)\) the set of inequality constraint indices that are active at a feasible solution x, that is, \(J\left( x\right) =\left\{ j\in J:g_{j}(x)=0\right\} .\)
Definition 5
A feasible point \(\overline{x}\) is said to be a weak Pareto (weakly efficient) solution of (MOP) if and only if there exists no feasible point x such that
Definition 6
A feasible point \(\overline{x}\) is said to be a Pareto (efficient) solution of (MOP) if and only if there exists no feasible point x such that
Let \(E:R^{n}\rightarrow R^{n}\) be a given one-to-one and onto operator. Throughout the paper, we shall assume that the functions constituting the considered problem (MOP) are E-differentiable at any feasible solution.
Now, for the considered problem (MOP), we define its associated differentiable vector optimization problem (VP\(_{E}\)) as follows:
We call the problem (VP\(_{E}\)) an E-vector optimization problem associated to (MOP). Let
be the set of all feasible solutions of (VP\(_{E}\)). Further, we denote by \(J_{E}\left( x\right)\) the set of inequality constraint indices that are active at a feasible solution \(x\in \Omega _{E}\), that is, \(J_{E}\left( x\right) =\left\{ j\in J:\left( g_{j}\,\circ\, E\right) (x)=0\right\}\).
Definition 7
A feasible point \(E(\overline{x})\) is said to be a weak E-Pareto (weakly E-efficient) solution of (MOP) if and only if there exists no feasible point E(x) such that
Definition 8
A feasible point \(E(\overline{x})\) is said to be an E-Pareto (E-efficient) solution of (MOP) if and only if there exists no feasible point E(x) such that
Lemma 1
[1] Let \(E:R^{n}\rightarrow R^{n}\) be a one-to-one and onto. Then \(E\left( \Omega _{E}\right) =\Omega\).
Lemma 2
[1] Let \(\overline{z}\in \Omega _{E}\) be a weak Pareto solution (a Pareto solution) of the E-vector optimization problem (VP\(_{E}\)). Then \(E\left( \overline{z}\right)\) is a weak Pareto solution (a Pareto solution) of the considered multiobjective programming problem (MOP).
Now, we give the Karush–Kuhn–Tucker necessary optimality conditions for a feasible solution \(\overline{x}\) to be a weak Pareto solution in (VP\(_{E}\)). These conditions are, at the same time, the so-called E-Karush–Kuhn–Tucker necessary optimality conditions for \(E\left( \overline{x}\right)\) to be a weak E-Pareto solution in (MOP).
Theorem 1
[21] (E-Karush–Kuhn–Tucker necessary optimality conditions). Let \(\overline{x}\in \Omega _{E}\) be a weak Pareto solution of the E-vector optimization problem (VP\(_{E}\)) (and, thus, \(E\left( \overline{x}\right)\) be a weak E-Pareto solution of the considered problem (MOP)). Further, f, g, h be E-differentiable at \(\overline{x}\) and the E-Guignard constraint qualification [21] be satisfied at \(\overline{x}\). Then there exist Lagrange multipliers \(\overline{\lambda}\in R^{p}\), \(\overline{\mu }\in R^{m}\), \(\overline{\xi }\in R^{q}\) such that
3 Mixed E-Duality
In this section, a vector mixed E-dual problem is defined for the considered E-differentiable problem (MOP) with inequality and equality constraints.
Let the index set J be partitioned into two disjoint subset \(J_{1}\) and \(J_{2}\) such that \(J=J_{1}\cup J_{2}\) and the index set T be partitioned into two disjoint subset \(T_{1}\) and \(T_{2}\) such that T \(=T_{1}\cup T_{2}\). Let \(J_{1}\) be an index set such that \(J_{1}=J\backslash J_{2}\) and \(J_{2}=J\backslash J_{1},\) let \(\mu _{J_{K}}g_{J_{K}}=\sum _{j\in J_{k}} \mu _{j}g_{j} ,\; k=1,2\) and, moreover, \(\left| J_{1}\right|\) and \(\left| J_{2}\right|\) denote the cardinality of the index sets \(J_{1}\) and \(J_{2}\), respectively. Further, let \(T_{1}\) be an index set such that \(T_{1}=T\backslash T_{2}\) and \(T_{2}=T\backslash T_{1},\) let \(\xi _{T_{K}}h_{T_{K}}=\sum _{t\in T_{k}} \xi _{t}h_{t},\; k=1,2\) and, moreover, \(\left| T_{1}\right|\) and \(\left| T_{2}\right|\) denote the cardinality of the index sets \(T_{1}\) and \(T_{2}\), respectively. Let us denote the set
Now, for the define E-differentiable problem (MOP), we introduce the definition of the scalar Lagrange function \(L:\Omega \times R_{+}^{p}\times R_{+}^{\left| J_{1}\right| }\times R^{\left| T_{1}\right| }\rightarrow R\) as follows
Further, let \(E:R^{n}\rightarrow R^{n}\) be a given one-to-one and onto operator. Further, let us define the following set
Now, we define the following vector mixed E-dual problem (VMD\(_{E}\)) for the considered E-differentiable problem (MOP):
where all functions are defined in the similar way as for the considered vector optimization problem (MOP). Further, let \(\Gamma _{E}\) denote the set of all feasible solutions of (VMD\(_{E}\)), that is,
Further, \(Y_{E}=\{y\in R^{n}:\left( {y},{\lambda },{\mu },{\xi }\right) \in \Gamma _{E}\}\). We call (VMD\(_{E}\)) the vector mixed E-dual problem for the E-differentiable multiobjective optimization problem (MOP).
Note that if set \(J_{1}=\varnothing\) and \(T_{1}=\varnothing\) in (VMD\(_{E}\) ), then we get a vector Mond-Weir E-dual problem for (MOP) [5] and, moreover, if we set \(J_{2}=\varnothing\) and \(T_{2}=\varnothing\) in (VMD\(_{E}\)), then we obtain a vector Wolfe E-dual problem for (MOP) (see, for example, [1, 24]).
Now, we shall prove several mixed duality results between E-vector optimization problems (VP\(_{E}\)) and (VMD\(_{E}\)) under (generalized) V-E-invexity assumptions. Then, we use these duality results in proving several mixed E-duality results between vector optimization problems (MOP) and (VMD\(_{E}\)).
Theorem 2
(Mixed weak duality between (VP\(_{E}\)) and (VMD\(_{E}\))). Let x and \(\left( {y},{\lambda },{\mu },{\xi }\right)\) be any feasible solutions of the problems (VP\(_{E}\)) and (VMD\(_{E}\)), respectively. Further, assume that at least one of the following hypotheses is fulfilled:
-
(a)
each objective function \(f_{i}\), \(i\in I\), is \(\alpha _{i}\)-E-invex with respect to \(\eta\) at y on \(\Omega _{E}\cup Y_{E}\), each constraint function \(g_{j}\), \(j\in J\), is \(\beta _{j}\)-E-invex with respect to \(\eta\) at y on \(\Omega _{E}\cup Y_{E}\), the functions \(h_{t}\), \(t\in T^{+}\left( E\left( {y}\right) \right) =\left\{ t\in T:{\xi }_{t}>0\right\}\) and the functions \(-h_{t}\), \(t\in T^{-}\left( E\left( {y}\right) \right) =\left\{ t\in T:{\xi }_{t}<0\right\}\), are \(\gamma _{t}\)-E-invex with respect to \(\eta\) at y on \(\Omega _{E}\cup Y_{E}\).
-
(b)
\(\left( f\circ E\right) (\cdot )+\bigg [{\mu }_{J_{1}}\left( g_{J_{1}}\circ\; E\right) (\cdot )+{\xi }_{T_{1}}\left( h_{T_{1}}\circ\; E\right) (\cdot )\bigg ]e\) is V-E-pseudo-invex at y on \(\Omega _{E}\cup Y_{E}\), \({\mu }_{J_{2}}\left( g_{J_{2}}\circ\; E\right) (\cdot )\) is \(\beta _{j}\)-E-quasi-invex at y on \(\Omega _{E}\cup Y_{E}\), \({\xi }_{T_{2}} \left( h_{T_{2}}\circ \;E\right) (\cdot )\) is \(\gamma _{t}\)-E-quasi-invex at y on \(\Omega _{E}\cup Y_{E}\).
Then
$$\begin{aligned} \left( f\,\circ\, E\right) (x)\nless \left( f\,\circ\, E\right) (y)+\bigg [{\mu }_{J_{1}}\left( g_{J_{1}}\circ \,E\right) (y)+{\xi }_{T_{1}}\left( h_{T_{1}}\circ \,E\right) (y)\bigg ]e\text {.} \end{aligned}$$(9)
Proof
Let x and \(\left( {y},{\lambda },{\mu },{\xi }\right)\) be any feasible solutions of the problems (VP\(_{E}\)) and (VMD\(_{E}\)), respectively.
The proof of this theorem under hypothesis A). By means of contradiction, suppose that
Thus,
Multiplying each inequality Eq. (10) by \(\lambda _{i}\) and then adding both sides of the resulting inequalities, we get
Since \(\sum _{i=1}^{p}{\lambda }_{i}=1\), the following inequality
holds. By \(x\in \Omega _{E}\) and \(\left( {y},{\lambda },{\mu },{\xi }\right) \in \Gamma _{E}\), we have
Combining Eqs. (11)−(14), we get
By assumption, \(\left( {y},{\lambda },{\mu },{\xi }\right)\) is feasible in (VMD\(_{E}\)). Hence,
Since \(x\in \Omega _{E}\) and \(y\in \Omega _{E},\) one has
Since the functions \(f_{i}\), \(i\in I\), \(g_{j}\), \(j\in J\), \(h_{t}\), \(t\in T^{+}\), and \(-h_{t},\) \(t\in T^{-}\), are V-E-invex at y on \(\Omega _{E}\cup Y_{E}\), by Definition 2, the following inequalities
hold, respectively. Combining Eqs. (12) and (18), we have
Since \(\alpha _{i}(E({x}),E(y))>0\), \(i=1,2,...,p,\) the above inequalities yield
Multiplying Eq. (23) by the corresponding Lagrange multipliers, we get that the following inequality
holds. Multiplying inequalities Eqs. (19)−(21) by the corresponding Lagrange multipliers, respectively, we obtain
Combining the above inequalities with Eqs. (16)−(17), we obtain that the inequalities
hold. Since \(\beta _{j}(E({x}),E(y))>0\), \(j=1,2,...,m,\) \(\gamma _{t}(E({x}),E(y))>0,\) \(t=1,2,...,q,\) the above inequalities yield, respectively
Adding both sides of the inequalities Eqs. (31)−(33) and (24), we obtain that the inequality
holds, contradicts the first constraint of the vector mixed E-dual problem (VMD\(_{E}\)). This means that the proof of the mixed weak duality theorem between the E-vector optimization problems (VP\(_{E}\)) and (VMD\(_{E}\)) is completed under hypothesis A).
The proof of this theorem under hypothesis B). We proceed by contradiction. Suppose, contrary to the result, that Eq. (10) holds. Since the function \(\left( f\;\circ\; E\right) (\cdot )+\bigg [{\mu }_{J_{1}}\left( g_{J_{1}}\circ \;E\right) (\cdot )+{\xi }_{T_{1}}\left( h_{T_{1}}\circ\; E\right) (\cdot )\bigg ]e\) is V-E-pseudo-invex at y on \(\Omega _{E}\cup Y_{E}\), by Definition 3, the inequality
holds. By \(x\in \Omega _{E}\) and \(\left( {y},{\lambda },{\mu },{\xi } \right) \in \Gamma _{E}\), it follows that the relations Eqs. (12)−(14) are fulfilled. Since \({\mu }_{J_{2}}\left( g_{J_{2}}\circ \;E\right) (\cdot )\) and \({ \xi }_{T_{2}}\left( h_{T_{2}}\circ \; E\right) (\cdot )\) are V-E-quasi-invex at y on \(\Omega _{E}\cup Y_{E}\), by the foregoing above relations, Definition 4 implies that the inequalities
hold, respectively. Combining Eqs. (35), (36) and (37), it follows that the inequality (34) is fulfilled, contradicting the first constraint of the vector mixed E-dual problem (VMD\(_{E}\)). This means that the proof of the mixed weak duality theorem between the E-vector optimization problems (VP\(_{E}\)) and (VMD\(_{E}\)) is completed under hypothesis B).
Theorem 3
(Mixed weak E-duality between (MOP) and (VMD\(_{E}\))). Let \(E\left( x\right)\) and \(\left( {y},{\lambda },{\mu },{\xi }\right)\) be any feasible solutions of the problems (MOP) and (VMD\(_{E}\)), respectively. Further, assume that all hypotheses of Theorem 2 are fulfilled. Then, mixed weak E-duality between (MOP) and (VMD\(_{E}\)) holds, that is,
Proof
Let \(E\left( x\right)\) and \(\left( {y},{\lambda },{\mu },{\xi }\right)\) be any feasible solutions of the problems (MOP) and (VMD\(_{E}\)), respectively. Then, by Lemma 1, it follows that x is any feasible solution of (VP\(_{E}\)). Since all hypotheses of Theorem 2 are fulfilled, the mixed weak E-duality theorem between the problems (MOP) and (VMD\(_{E}\)) follows directly from Theorem 2.
If some stronger V-E-invexity hypotheses are imposed on the functions constituting the considered E-differentiable problem, then the following result is true.
Theorem 4
(Mixed weak duality between (VP\(_{E}\)) and (VMD\(_{E}\))). Let x and \(\left( {y},{\lambda },{\mu },{\xi }\right)\) be any feasible solutions of the problems (VP\(_{E}\)) and (VMD\(_{E}\)), respectively. Further, assume that at least one of the following hypotheses is fulfilled:
-
(A)
each objective function \(f_{i}\), \(i\in I\), is strictly \(\alpha _{i}\)-E-invex with respect to \(\eta\) at y on \(\Omega _{E}\cup Y_{E}\), each constraint function \(g_{j}\), \(j\in J\), is \(\beta _{j}\)-E-invex with respect to \(\eta\) at y on \(\Omega _{E}\cup Y_{E}\), the functions \(h_{t}\), \(t\in T^{+}\left( E\left( {y}\right) \right)\) and the functions \(-h_{t}\), \(t\in T^{-}\left( E\left( {y}\right) \right)\), are \(\gamma _{t}\)-E-invex with respect to \(\eta\) at y on \(\Omega _{E}\cup Y_{E}\).
-
(B)
\(\left( f\;\circ\; E\right) (\cdot )+\bigg [{\mu }_{J_{1}}\left( g_{J_{1}}\;\circ \;E\right) (\cdot )+{\xi }_{T_{1}}\left( h_{T_{1}}\;\circ \;E\right) (\cdot )\bigg ]e\) is strictly V-E-pseudo-invex at y on \(\Omega _{E}\cup Y_{E}\), \({\mu }_{J_{2}}\left( g_{J_{2}}\;\circ \;E\right) (\cdot )\) is \(\beta _{j}\)-E-quasi-invex at y on \(\Omega _{E}\cup Y_{E}\), \({\xi }_{T_{2}} \left( h_{T_{2}}\;\circ\; E\right) (\cdot )\) is \(\gamma _{t}\)-E-quasi-invex at y on \(\Omega _{E}\cup Y_{E}\).
Then
$$\begin{aligned} \left( f\;\circ \;E\right) (x)\nleq \left( f\;\circ\; E\right) (y)+\bigg [{\mu }_{J_{1}}\left( g_{J_{1}}\;\circ\; E\right) (y)+{\xi }_{T_{1}}\left( h_{T_{1}}\circ \;E\right) (y)\bigg ]e. \end{aligned}$$(38)
Theorem 5
(Mixed weak E-duality between (MOP) and (VMD\(_{E}\))). Let \(E\left( x\right)\) and \(\left( {y},{\lambda },{\mu },{\xi }\right)\) be any feasible solutions of the problems (MOP) and (VMD\(_{E}\)), respectively. Further, assume that all hypotheses of Theorem 4 are fulfilled. Then, mixed weak E-duality between (MOP) and (VMD\(_{E}\)) holds, that is,
Theorem 6
(Mixed strong duality between (VP\(_{E}\)) and (VMD\(_{E}\)) and also strong E-duality between (MOP) and (VMD\(_{E}\))). Let \(\overline{x}\in \Omega _{E}\) be a weak Pareto solution (a Pareto solution) of the E-vector optimization problem (VP\(_{E}\)) and the E-Guignard constraint qualification [21] be satisfied at \(\overline{x}.\) Then there exist \(\overline{\lambda }\in R^{p}\), \(\overline{\lambda }\ne 0\), \(\overline{\mu }\in R^{m}\), \(\overline{\mu }\geqq 0\), \(\overline{\xi }\in R^{q}\), \(\overline{\xi }\geqq 0\) such that \(\left( \overline{x},{\overline{\lambda } },{\overline{\mu } },{\overline{\xi } }\right)\) is feasible for the problem (VMD\(_{E}\)) and the objective functions of (VP\(_{E}\)) and (VMD\(_{E}\)) are equal at these points. If also weak duality (Theorem 2) holds between (VP\(_{E}\)) and (VMD\(_{E}\)), then \(\left( \overline{x},{\overline{\lambda } },{\overline{\mu } },{\overline{\xi } }\right)\) is a (weak) maximum point for (VMD\(_{E}\)).In other words, if \(E(\overline{x})\in \Omega\) is a (weak) E-Pareto solution of the problem (MOP), then \(\left( \overline{x},{\overline{\lambda } },{\overline{\mu } },{\overline{\xi } }\right)\) is a (weak) efficient solution of a maximum type in the vector mixed dual problem (VMD\(_{E}\)). This means that the mixed strong E-duality holds between the problems (MOP) and (VMD\(_{E}\)).
Proof
Since \(\overline{x}\in \Omega _{E}\) is a (weak) Pareto solution of the problem (VP\(_{E}\)) and the E-Guignard constraint qualification [21] is satisfied at \(\overline{x}\), by Theorem 1, there exist \(\overline{\lambda }\in R^{p}\), \(\overline{\lambda }\ne 0\), \(\overline{\mu }\in R^{m}\), \(\overline{\mu }\geqq 0\), \(\overline{\xi }\in R^{q}\), \(\overline{\xi }\geqq 0\) such that \(\left( \overline{x},{\overline{\lambda } },{\overline{\mu } },{\overline{\xi } }\right)\) is a feasible solution of the problem (VMD\(_{E}\)). This means that the objective functions of (VP\(_{E}\)) and (VMD\(_{E}\)) are equal. If we assume that weak duality (Theorem 2) holds between (VP\(_{E}\)) and (VMD\(_{E}\)), \(\left( \overline{x},{\overline{\lambda } },{\overline{\mu } },{\overline{\xi } }\right)\) is a (weak) maximum point for (VMD\(_{E}\)) in the sense of mixed.
Moreover, we have, by Lemma 1, that \(E\left( \overline{x}\right) \in \Omega\). Since \(\overline{x} \in {\Omega}_E\) is a weak Pareto solution of the problem (VP\(_{E}\)), by Lemma 2, it follows that \(E\left( \overline{x}\right)\) is a weak E-Pareto solution in the problem (MOP). Then, by the strong duality between (VP\(_{E}\)) and (VMD\(_{E}\)), we conclude that also the mixed strong E-duality holds between the problems (MOP) and (VMD\(_{E}\)). This means that if \(E\left( \overline{x}\right) \in \Omega\) is a weak E-Pareto solution of the pr`oblem (MOP), there exist \(\overline{\lambda }\in R^{p}\), \(\overline{\lambda }\ne 0\), \(\overline{\mu }\in R^{m}\), \(\overline{\mu }\geqq 0\), \(\overline{\xi }\in R^{q}\), \(\overline{\xi }\geqq 0\) such that \(\left( \overline{x},{\overline{\lambda } },{\overline{\mu } },{\overline{\xi } }\right)\) is a weakly efficient solution of a maximum type in the mixed dual problem (VMD\(_{E}\)).
Theorem 7
(Mixed converse duality between (VP\(_{E}\)) and (VMD\(_{E}\))). Let \(\left( { \overline{x}},{\overline{\lambda }},{\overline{\mu }},{\overline{\xi }} \right)\) be a (weakly) efficient solution of a maximum type in mixed E-dual problem (VMD\(_{E}\)) such that \(\overline{x}\in \Omega _{E}\). Moreover, that the objective functions \(f_{i}\), \(i\in I\), are \(\alpha _{i}\)-E-invex with respect to \(\eta\) at \({\overline{x}}\) on \(\Omega _{E}\cup Y_{E}\), the constraint functions \(g_{j}\), \(j\in J\), are \(\beta _{j}\)-E-invex with respect to \(\eta\) at \({\overline{x}}\) on \(\Omega _{E}\cup Y_{E}\), the functions \(h_{t}\), \(t\in T^{+}\left( E\left( {\overline{ x}}\right) \right)\) and the functions \(-h_{t}\), \(t\in T^{-}\left( E\left( { \overline{x}}\right) \right)\), are \(\gamma _{t}\)-E-invex with respect to \(\eta\) at \(\overline{x}\) on \(\Omega _{E}\cup Y_{E}\). Then \(\overline{x}\) is a (weak) Pareto solution of the problem (VP\(_{E}\)).
Proof
Let \(\left( {\overline{x}},{\overline{\lambda }},{\overline{\mu }},{\overline{\xi }} \right)\) be a (weakly) efficient solution of a maximum type in mixed E-dual problem (VMD\(_{E}\)) such that \(\overline{x}\in \Omega _{E}\). By contradiction, suppose that
Thus,
Multiplying each inequality Eq. (39) by \(\overline{\lambda } _{i}\) and then adding both sides of the resulting inequalities, we get
Since \(\sum _{i=1}^{p}{\overline{\lambda } }_{i}=1\), the following inequality
holds. By \(x\in \Omega _{E}\) and \(\left( {\overline{x}},{\overline{\lambda } },{\overline{\mu }},{\overline{\xi } }\right) \in \Gamma _{E}\), we have
Combining Eqs. (40)−(43), we get
By assumption, \(\left( {\overline{x}},{\overline{\lambda } },{\overline{\mu } },{\overline{\xi } }\right)\) is feasible in (VMD\(_{E}\)). Hence,
Since \(x\in \Omega _{E}\) and \(\overline{x}\in \Omega _{E},\) one has
Since the functions \(f_{i}\), \(i\in I\), \(g_{j}\), \(j\in J\), \(h_{t}\), \(t\in T^{+}\), and \(-h_{t},\) \(t\in T^{-}\), are V-E-invex at \(\overline{x}\) on \(\Omega _{E}\cup Y_{E}\), by Definition 2, the following inequalities
hold, respectively. Combining Eqs. (41) and (47), we have
Since \(\alpha _{i}(E({x}),E(\overline{x}))>0\), \(i=1,2,...,p,\) the above inequalities yield
Multiplying Eq. (52) by the corresponding Lagrange multipliers, we get that the following inequality
holds. Multiplying inequalities Eqs. (48)−(50) by the corresponding Lagrange multipliers, respectively, we obtain
Combining the above inequalities with Eqs. (45)−(46), we obtain that the inequalities
hold. Since \(\beta _{j}(E({x}),E(\overline{x}))>0\), \(j=1,2,...,m,\) \(\gamma _{t}(E({x}),E(\overline{x}))>0,\) \(t=1,2,...,q,\) the above inequalities yield, respectively
Adding both sides of the inequalities Eqs. (60)−(62) and Eq. (53), we obtain that the inequality
holds, contradicts the first constraint of the vector mixed E-dual problem (VMD\(_{E}\)). This means that the proof of this theorem is completed.
Theorem 8
(Mixed converse E-duality between (MOP) and (VMD\(_{E}\))). Let \(\left( {\overline{x}},{\overline{\lambda }},{\overline{\mu }},{\overline{\xi }}\right)\) be a (weakly) efficient solution of a maximum type in mixed dual problem (VMD\(_{E}\)). Further, assume that all hypotheses of Theorem 7 are fulfilled. Then \(E(\overline{x})\in \Omega\) is a (weak) E-Pareto solution of the problem (MOP).
Proof
The proof of this theorem follows directly from Lemma 2 and Theorem 7.
Theorem 9
(Mixed restricted converse duality between (VP\(_{E}\)) and (VMD\(_{E}\))). Let \(\overline{x}\) and \(\left( {\overline{y}},{\overline{\lambda }},{\overline{\mu }},{\overline{\xi }}\right)\) be feasible solutions for the problems (VP\(_{E}\)) and (VMD\(_{E}\)), respectively, such that
Moreover, assume that \(\overline{\lambda }_{i} \kern 0.03000 em f_{i}\), \(i\in I\), are \(\alpha _{i}\)-E-pseudo-invex at \({\overline{y}}\) on \(\Omega _{E}\cup Y_{E}\), the functions \(\overline{\mu }_{J_{2}}\left( g_{J_{2}}\circ\; E\right) (\cdot )\) is \(\beta _{j}\)-E-quasi-invex at \(\overline{y}\) on \(\Omega _{E}\cup Y_{E}\), \(\overline{\xi }_{T_{2}} \left( h_{T_{2}}\circ \; E\right) (\cdot )\) is \(\gamma _{t}\)-E-quasi-invex at \(\overline{y}\) on \(\Omega _{E}\cup Y_{E}\). Then \(\overline{x}=\overline{y}\), that is, \(\overline{x}\) is a (weak) Pareto solution of the problem (VP\(_{E}\)) and \(\left( {\overline{y}},{\overline{\lambda }},{\overline{\mu }},{\overline{\xi }}\right)\) is a (weakly) efficient point of maximum type for the problem (VMD\(_{E}\)).
Proof
The proof of this theorem follows directly from Theorem 2 and Theorem 4.
Theorem 10
(Mixed restricted converse E-duality between (MOP) and (VMD\(_{E}\))). Let \(\left( {\overline{y}},{\overline{\lambda }},{\overline{\mu }},{\overline{\xi }}\right)\) be a feasible solution of the problem (VMD\(_{E}\)). Further, assume that there exist \(E\left( \overline{x}\right) \in \Omega\) such that \(\overline{x}=\overline{y}\). If all hypotheses of Theorem 9 are fulfilled, then \(E\left( \overline{x}\right)\) is an E-Pareto solution of the problem (MOP) and \(\left( {\overline{y}},{\overline{\lambda }},{\overline{\mu }},{\overline{\xi }}\right)\) is a weakly efficient solution of maximum type for the problem (VMD\(_{E}\)).
Proof
The proof of this theorem follows directly from Lemma 2 and Theorem 9.
4 Conclusion
This paper analyzes mixed E-duality results for E-differentiable V-E-invex multiobjective programming problems with both inequality and equality constraints. The so-called vector mixed E-dual problem has been formulated for such nonconvex (not necessarily)differentiable multiobjective programming problems. Then, various mixed E-duality theorems between the considered E-differentiable vector optimization problem and its mixed dual problem have been proved under (generalized) V-E-invexity hypotheses. The results established in this paper for E-differentiable vector optimization problems extend and generalize similar duality results in the sense of mixed established under other concepts of E-differentiable (generalized) convexity and also duality results in the sense of Mond-Weir and in the sense of Wolfe established for such multicriteria optimization problems.
However, some interesting topics for further research remain. It would be of interest to investigate whether it is possible to prove similar results for other classes of E-differentiable vector optimization problems. We shall investigate these questions in subsequent papers.
References
Antczak T, Abdulaleem N (2019) E-optimality conditions and Wolfe E-duality for E-differentiable vector optimization problems with inequality and equality constraints. J Nonlinear Sci Appl 12:745–764
Ahmad I, Gulati TR (2005) Mixed type duality for multiobjective variational problems with generalized (F, ρ)-convexity. J Math Anal Appl 306(2):669–683
Ahmad I (2005) Multiobjective mixed symmetric duality with invexity. New Zealand J Math 34(1):1–9
Abdulaleem N (2021) V-E-invexity in E-differentiable multiobjective programming. Numer Algebra Control Optim. https://doi.org/10.3934/naco.2021014
Abdulaleem N (2019) E-duality results for E-differentiable vector optimization problems under (generalized) E-convexity. 6th Intl Confe Re Adv in Pure and Appl Math, Istanbul, 9-26
Abdulaleem N (2019) E-duality results for E-differentiable E-invex multiobjective programming problems. In Journal of Physics: Conference Series (Vol. 1294, No. 3, p. 032027). IOP Publishing
Bector CR, Chandra S (2001) On mixed duality in mathematical programming. J Math Anal Appl 259:346–356
Bot RI, Grad SM, Wanka G (2009) Duality in vector optimization. Springer Science & Business Media
Cambini R, Carosi L (2010) Mixed Type Duality for Multiobjective Optimization Problems with Set Constraints, in Optimality Conditions in Vector Optimization, M. A. Jimenéz, G. Ruiz-Garzón and A. Rufián-Lizana eds., Bentham Science Publishers, 119-142
Habibi S, Kanzi N, Ebadian A (2020) Weak slater qualification for nonconvex multiobjective semi-infinite programming. Iran J Sci Technol Trans A Sci: Science 44(2):417–424
Jahn J (1983) Duality in vector optimization. Math Program 25(3):343–353
Jayswal A, Jha S, Prasad AK, Ahmad I (2018) Second-order symmetric duality in variational control problems over cone constraints. Asia Pac J Oper Res 35(4):1850028
Kanzi N, Shaker AJ, Caristi G (2018) Optimality, scalarization and duality in linear vector semi-infinite programming. Optimization 67(5):523–536
Mukherjee RN, Rao CP (2000) Mixed type duality for multiobjective variational problems. J Math Anal Appl 252(2):571–586
Piao GR (2015) Jiao L (2015) Optimality and mixed duality in multiobjective E-convex programming. J Inequal Appl 1:1–13
Treanţă S, Mititelu Ş (2019) Duality with (ρ, b)-quasiinvexity for multidimensional vector fractional control problems. J Inf Optim Sci 40(7):1429–1445
Treanţă S, Mititelu Ş (2020) Efficiency for variational control problems on Riemann manifolds with geodesic quasiinvex curvilinear integral functionals. Revista de la Real Academia de Ciencias Exactas, Físicas y Naturales. Serie A. Matemáticas 114(3):1-15
Xu Z (1996) Mixed type duality in multiobjective programming problems. J Math Anal Appl 198(3):621–635
Youness EA (1999) E-convex sets, E-convex functions, and E-convex programming. J Optim Theory Appl 102:439–450
Megahed AA, Gomma HG, Youness EA, El-Banna AZ (2013) Optimality conditions of E-convex programming for an E-differentiable function. J Inequal Appl 2013:246
Abdulaleem N (2019) E-invexity and generalized E-invexity in E-differentiable multiobjective programming, In: ITM Web of Conferences, (Vol. 24, p. 01002), EDP Sciences
Hanson MA (1981) On sufficiency of the Kuhn-Tucker conditions. J Math Anal Appl 80(2):545–550
Jeyakumar V, Mond B (1992) On generalised convex mathematical programming. Anziam J 34(1):43–53
Abdulaleem N (2018) Wolfe E-duality for E-differentiable E-invex vector optimization problems with inequality and equality constraints. 2018 Intl Conf on Applied Math & Comput Sci (ICAMCS.NET), Budapest, Hungary, 156-163
Acknowledgements
The author expresses his deep thanks and his respect to the referees and the Journal of SN Operations Research Forum for the valuable comments in the evaluation of this paper.
Author information
Authors and Affiliations
Ethics declarations
Competing Interests
The author declares to have no competing interests.
Additional information
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
This article is part of the Topical Collection on Research Developments in Theory and Algorithms of Numerical Analysis and Optimization: Proceedings from the Fifth Conference on Numerical Analysis and Optimization — NAOV 2020
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Abdulaleem, N. Mixed E-duality for E-differentiable Vector Optimization Problems Under (Generalized) V-E-invexity. SN Oper. Res. Forum 2, 32 (2021). https://doi.org/10.1007/s43069-021-00074-z
Received:
Accepted:
Published:
DOI: https://doi.org/10.1007/s43069-021-00074-z