Generalised variational-like inequalities and a gap function (Q4243348)

Let \(H\) be a Hilbert space with and let \(K\) be any subset of \(H\). Recall that for all \(u,v\in K\) and \(t\in[0,1]\), the set \(K\) is said to be an invex set, if there exists a function \(\eta(.,.):K\times K\to H\) such that \(u+t\eta(v, u)\in K\). \(K\) is said to be an invex set with respect to \(\eta(.,.)\) if \(K\) is invex at each \(u\in K\). The function \(F: K\to H\) is said to be preinvex with respect to \(\eta(.,.)\), if for all \(u,v\in K\) and \(t\in[0,1]\), \[ F(u+ t\eta(u))\leq (1- t)F(u)+ tF(v). \] It is clear that for \(\eta(v,u)= v-u\) the preinvex function is a convex function and the invex set is a convex set. However, there are examples that a preinvex function is not a convex function and the invex set is not a convex set. It has been showen [\textit{M. A. Noor}, Optimization 30, No. 4, 323-330 (1994; Zbl 0816.49005)] that the minimum of a differentiable preinvex function on an invex set can be characterized by an inequality of the type: \[ \langle F'(u), \eta(v,u)\rangle\geq 0.\tag{1} \] An inequality of this type is known as the variational-like inequality. Note that for \(\eta(n, v)=v-u\), the problem (1) is called the variational inequality problem. We would like to point out that the variational-like inequalities are generalization of the classical variational inequalities, which are studied in the context of the convexity theory, whereas the variational-like inequalities are studied in the setting of the \textit{invexity} theory. In the study of the variational-like inequalities, the invexity plays an important part. Unfortunately, up to now, except for a few cases, all the results are obtained for variational-like inequalities in the context of convexity. No attention has been paid to the invexity role. Consequently, all these results are trivial and artificial extensions of the previously known results. In the paper under review, the authors have considered the problems of the variational-like inequalities of the type: Find \(\overline x\in C\) and \(\overline s\in T(\overline x)\) such that \[ \langle\theta(\overline x,\overline s),\;\eta(\overline x,y)\rangle> 0,\quad \text{for all }y\in C \] in the setting of the convexity theory. In view of the above remarks, all results obtained in this paper are trivial, straight forward and meaningless generalizations of the previously known results.