Modified two-step approach to the solution of constrained nonlinear optimisation problems

The basic idea of the modified two-step approach in optimising control is adopted to design an algorithm for the solution of constrained nonlinear optimisation problems. The algorithm is less sensitive to the Hessian structure of the problem than many existing optimisation algorithms and, hence, more effective for solving ill-conditioned problems. Optimality and convergence analyses are provided. Practical issues concerning algorithm implementation, such as model based optimisation, model matrix and iterative gain selection, are discussed and computer simulations show that the algorithm is very efficient.<
>