No label defined (Q4528089)

This article deals with the solvability of the following variational inequality \[ \bigl\langle f(u_*), x-x_*\bigr\rangle \geq 0\quad (u_*\in Tx_*,x\in K), \] where \(f:H\to H\), \(T:H\to P(H)\) is a multivalued mapping, \(H\) a Hilbert space, \(K\) a nonempty closed convex subset of \(H\). The main part of the article is devoted to the analysis of algorithms of approximative solving of the inequality; the corresponding procedure is defined by formulas \[ \bigl\langle f(u_k)+ x_{k+1}-y_k, x-x_{k+1} \bigr\rangle \geq 0\;\bigl(u_k \in T(y_k), x\in K\bigr), \] \[ \bigl\langle f(v_k)+ y_k-x_k,x-y_k \bigr\rangle \geq 0\;\bigl(v_k\in T(x_k),x\in K). \] The convergence of these approximations is proved; the simple error estimates are also given.