A note on the regularity of reduced models obtained by nonlocal quasi-continuum-like approaches (Q1016119)

The authors consider the optimization problem \[ \min_{x_1,x_2}\,f(x_1, x_2)\text{ s.t. }g_1(x_1)= 0,\quad g_2(x_2)= 0,\quad g_3(x_1, x_2)= 0, \] where \(g_1(x_1): \mathbb{R}^m\to \mathbb{R}^{q_1}\), \(g_2(x_2): \mathbb{R}^{n-m}\to \mathbb{R}^{q_2}\) and \(g_3(x_1, x_2): \mathbb{R}^n\to \mathbb{R}^{q_3}\) are the constraint functions, which, together with the objective function \(f(x_1,x_2): \mathbb{R}^n\to \mathbb{R}\), are twice continuously differentiable. For these problems, model reduction techniques are investigated that are based on a nonlocal quasi-continuum-like approach. These techniques reduce a large optimization problem to either a system of nonlinear equations or another optimization problem that are expressed in a smaller number of degree of freedom. Numerical experiments are given.