A double projection algorithm for multi-valued variational inequalities and a unified framework of the method (Q555340)

The goal of this paper is to find the solution \(x^* \in C\) and \(\xi \in F(x^*)\) of the generalized variational inequality such that \[ \langle \xi, y-x^* \rangle \geq 0, \forall y \in C\tag{1} \] where \(C\) is a nonempty closed convex set in \(\mathbb{R}^n\), \(F\) is a multi-valued mapping from \(C\) into \(\mathbb{R}^n\) with nonempty values, and \(\langle .,.\rangle\) and \(\|.\|\) denote the inner product and norm in \(\mathbb{R}^n\), respectively. The authors assume that the solution set \(S\) of the problem (1) is nonempty and \(F\) is continuous on \(C\) with nonempty compact convex values satisfies the inequality \[ \langle \zeta, y-x\rangle \geq 0, \forall y \in C, \zeta \in F(y), \forall x \in S.\tag{2} \] (Here property (2) holds if \(F\) is pseudomonotone on C in the sense of Karamardian. In particular, if is monotone, then (2) holds.) Main result: {\parindent6mm \begin{itemize}\item[1)] If \(F:C \rightarrow 2^{R_n}\) is continuous with nonempty compact convex values on \(C\) and the condition (2) holds, then either the proposed algorithm terminates in a finite number of iterations or generates an infinite sequence \(\{x_i\}\) converging to a solution of (1). \item[2)] If \(F\) is also Lipschitz continuous with modulus \(L > 0\) and if there exist positive constants \( \langle c, \delta \rangle \) such that \(dist (\mathbf{x}, S) \leq c \|r_\mu (\mathbf{x}, \xi)\|\), \(\forall (\mathbf{x}, \xi) \in P(\delta)\), then there is a constant \(\alpha > 0\) such that for sufficiently large \(i\) one has: \(dist(x_i, s) \leq (1/(\alpha i + dist^{-2}(x_0, s)))^{1/2}\). \end{itemize}} The details of the algorithm are presented and several preliminary results for convergence analysis are proved. The authors give a unified framework of a projection-type algorithm for multi-valued variational inequalities. At the end the results of some numerical experiments for the proposed algorithm are presented.