PŁ-like gradient-mapping relation for trust-region optimisation
Let EEE be the objective function, let s(k)\boldsymbol{s}^{(k)}s(k) be the iterate at iteration kkk, let E∗E^*E∗ denote the optimal objective value, and let…