Proximal point method and elliptic regularization (Q838049)

The authors develop a generalized proximal point method for solving variational inequalities with maximal monotone operators. It admits a successive approximation of the feasible set and of a symmetric component of the operator as well as an inexact solution of the regularized problem. The conditions on the approximation are coordinated with the properties of finite element methods for solving problems in mathematical physics. The choice of the regularizing functional exploits a possible ``reserve of monotonicity'' of the operator in the variational inequality. In the special cases of the minimal surface problem and related variational inequalities, as well as for the convection-diffusion problem, the developed method extends the principle of elliptic regularization. A special convergence analysis shows a more qualitative convergence of the method applied to these problems than it follows from the general theory of proximal point methods. Applications to some variational inequalities from elasticity theory are investigated as well.