Abstract: A new projection algorithm is proposed for mixed variational inequalities. Mixed variational inequalities are more general than classical variational inequalities and have rather practical applications in elastoplastic problems. Assuming that the underlying mapping F is pseudo-monotone with respect
to the underlying function $\varphi$, the author proves that this new method is globally convergent. In addition, when a certain error bound holds locally, the author analyzes the convergence rate of the iteration sequence.

Key words: Mixed variational inequalities, Projection algorithm, Elastoplastic problems, Pseudomonotone