The convergence of estimators based on heuristics: theory and application to a GARCH model (Q964667)

A threshold accepting optimization algorithm is implemented for calculation of maximum likelihood (ML) estimators of the parameters of GARCH(1,1) models. The performance of the algorithm is assessed via simulations. It is compared to the ML estimator of the GARCH package for MathLab. The authors' conclusion is that their algorithm outperforms the MathLab in most cases.