Global optimization for identification

The formulation of the transfer function identification problem leads directly to a nonlinear optimization problem. This nonlinear optimization problem is non-convex and may exhibit many local optima. As a result of the presence of local optimum, optimization methods based upon gradient techniques cannot be guaranteed to converge to the global optimum. A relaxation branch and bound technique is proposed to solve the problem. The algorithm is presented and its convergence properties discussed. Finally, several simulation examples utilizing the global technique are provided