An algorithm for smooth nonlinear constrained optimization problems is described, in which a sequence of feasible iterates is generated by solving a trust-region sequential quadratic programming (SQP) subproblem at each iteration and by perturbing the resulting step to retain feasibility of each iterate. By retaining feasibility, the algorithm avoids several complications of other trust-region SQP approaches: the objective function can be used as a merit function, and the SQP subproblems are feasible for all choices of the trust-region radius. Global convergence properties are analyzed under various assumptions on the approximate Hessian. Under additional assumptions, superlinear convergence to points satisfying second-order sufficient conditions is proved.
No takes yet. Share an insight, caveat, or question.
Wright et al. (2004) studied this question.
Synapse has enriched 3 closely related papers on similar clinical questions. Consider them for comparative context: