A class of nonmonotone trust region algorithms for constrained optimizations (Q1906603)

The authors consider the problem of minimizing a smooth function \(f\) over a closed convex set \(\Omega\). They state a trust region algorithm that includes a variant of the nonmonotone line search idea. It is stated without proof that under reasonable assumptions, at least one limit-point of the sequence generated by the algorithm satisfies the first-order optimality system.