An implementable augmented Lagrange method for solving fixed point problems with coupled constraints (Q629292): Difference between revisions

Let \(\Omega_0\subset{\mathbb R}^n\) be closed and convex and let \(\Phi:{\mathbb R}^n\times{\mathbb R}^n\to{\mathbb R}\) and \(g:{\mathbb R}^n\times{\mathbb R}^n\to{\mathbb R}^m\) be given maps such that \(\Phi(v,\cdot)\) and the \(m\) components of \(g(v,\cdot)\) are convex functions in \(\Omega_0\) for each \(v\in\Omega_0\). This paper is devoted to an ``augmented Lagrange'' procedure for the problem of finding \(v^*\in\Omega_0\) such that \[ v^*\in\text{arg~min}\big\{\Phi(v^*,\omega)|g(v^*,\omega)\leq 0, \omega\in\Omega_0\big\}. \tag{1} \] In the proposed procedure the semismooth Newton method is used at each step to solve the inner implicit problems. Under some further assumptions the authors prove that their method converges monotonically in norm to a solution of (1). The paper contains some numerical examples illustrating the proposed method.