Generalized iterative process and associated regularization for \(J\)-pseudomonotone mixed variational inequalities (Q1030211)

The main purpose of this paper is to study a generalized iterative process and associated regularization for \(J\)-pseudomonotone mixed variational inequalities. A generalized iterative process for solving a class of variational inequalities (the mixed variational inequalities) is proposed. The author proves the weak convergence of the entire generalized iterative sequence in a \(p\)-uniformly smooth Banach space, \(p>1\) which admits a weakly sequentially continuous duality mapping. The general results are applied to the stationary filtration of an incompressible fluid governed by a discontinuous filtration law with limit gradient. The convergence of an iterative regularization method for the variational inequalities (with \(J\)-pseudomonotone operators in uniformly smooth Banach spaces under the assumption that the iterative parameter may vary from step to step is analyzed. Main result: An efficient face representation and recognition method (which combines 2DPCA with 2DLDA) is proposed. The major advantage of the proposed method is to combine the optimal reconstruction information from 2DPCA and the optimal discriminative information form 2DLDA. Finally experimental results reveal that all methods perform well under the condition of minor variation of tilting, scaling and rotation (on ORL database). The proposed method has better recognition accuracy under the condition of varied facial expression and lighting configuration (on Yale database). Thus the author's method is more robust.