“…These include deterministic, gradient-based methods such as Gauss-Newton or Gauss-Marquardt (e.g., [61,89,100,101]), reduced-gradient methods [10,13,28,51,72,86], and successive quadratic programming, which is the most used among industrial DR works, as well as derivative-free, metaheuristic random search algorithms, such as particle swarm optimization (PSO) [18,64]. Applicable deterministic solvers for large-scale nonlinear optimization problems are available as commercial and noncommercial software [48,[125][126][127], and generally incorporate second-derivative information from the optimization model, exploit the sparsity of the KKT matrix and problem structure, deal efficiently with large sets of active constraints, and handle dependent constraints and negative curvature.…”