Vector optimization problems with nonconvex preferences (Q2481362)

This paper studies vector optimization problems where the order is defined by a cone, which is not necessarily convex. First, the authors study six types of variational inequalities. For example, let \(X\) and \(Y\) be two Banach spaces ordered by cones \(C\subset X\) and \(P\subset Y\), respectively. The cone \(C\) is convex, whereas \(P\) is not. Let \(T\) be a mapping from \(X\) to the set of bounded linear operators from \(X\) to \(Y\), \(L(X,Y)\). Then a {\textit{strong vector variational inequality problem}} is defined as finding \(x\in C\) such that and \((Tx,y-x)\geq_P 0\) for every \(y\in C\). A problem of vector optimization of functional \((Tx,x)\) subject to a system of linear inequalities is called either strong or mild depending on whether the constraints are formulated in term of the cone \(P\) or its complement, but with the opposite sign of the inequality. A series of theorems formulate various conditions for existence of solution, as well as necessary optimality conditions in terms of variational inequalities. In Conclusions, the authors state that the results of this work will help to develop algorithms for finding the whole sets of solutions to vector optimization problems with nonconvex ordering cones. This statement is not supported by a single example in the paper, though.