A nonmonotone filter trust region method for nonlinear constrained optimization (Q953386)

The following minimization problem is considered: \[ \text{Minimize }f(x)\text{ subject to }c_i(x)=0\text{ for }i\in\{1,\dots,m\}, \] where \(x\in \mathbb{R}^n\), \(f:\mathbb{R}^n\to R\), and \(c_i(x)\), \(i\in\{1,\dots,m\}\) are twice continuously differentiable functions. The non-monotone filter technique and the fraction of Cauchy decrease are introduced. Using these concepts, the authors propose a non-monotone filter trust region algorithm for solving the minimization problem. Convergence properties and some numerical results in the concluding part of the paper show the efficiency of the proposed algorithm.