A new extragradient method for single-valued variational inequality (Q2863760)

The authors consider the variational inequality problem (VIP) which is to find a vector \(x^* \in C\) such that:NEWLINENEWLINE(1) \( \langle F(x^*), x-x^* \rangle \geq 0,\, \forall x \in C\),NEWLINENEWLINEwhere \(C\) is a nonempty closed convex set in \(\mathbb R^n\), \(F\) is a single-valued mapping from \(\mathbb R^n\) into itself, and \(\langle \cdot,\cdot\rangle\) and \(\|\cdot\|\) denote the inner product and norm in \(\mathbb R^n\), respectively. It is assumed that the solution set \(S\) of the problem (1) is nonempty and \(F\) is pseudomonotone on \(C\) with respect to the solution set \(S\), i.e.NEWLINENEWLINE(2) \( \langle F(y), y-x \rangle \geq 0,\, \forall y \in C,\, \forall x \in S\),NEWLINENEWLINEMain results: If \(\{x_i\}\) is the sequence generated by the proposed algorithm and \(x^* \in S\), and the assumption (2) holds, then one hasNEWLINENEWLINE(3) \(\|x_{i+1}-x^*\|^2 \leq \|x_i - x^*\|^2-(1-\sigma^2)\rho^2_i\|r_1(x_i)\|^2\).NEWLINENEWLINEIf \(F: C \rightarrow \mathbb R^n\) is continuous on \(C\) and the assumption (2) holds, then the sequence \(\{x_i\}\) generated by the given algorithm converges to a solution \(\overline{x}\) of (1).NEWLINENEWLINEIf \(F\) is Lipschitz continuous with modulus \(L>0\) and if there exist positive constants \(c\) and \(\delta\) such that \(\mathrm{dist} (x,S) \leq c \|r_1(x)\|\), for all \(x\) satisfying \(\|r_1(x)\| \leq \delta\), then any sequence \(\{x_i\}\) generated by the presented algorithm converges strongly to a solution \(\overline{x}\) of (1) and the rate of convergence is \(\mathbb R\)-linear.NEWLINENEWLINEThe authors introduce an extra-gradient algorithm for the VIP and obtain a global convergence theorem, assuming that \(F\) is continuous on \(C\).