The paper proposes a constrained Nonlinear Programming methodology for volatility estimation with GARCH models. These models are usually developed and solved as unconstrained optimization problems whereas they actually fit into nonlinear, nonconvex problems. Computational results on FTSE 100 and S & P 500 indices with up to 1500 data points are given and contrasted to Diagonal VECH and BEKK models from the S-Plus library
Bilkent University, Department of Industrial Engineering Technical Report, July 2001.
View Constrained Nonlinear Programming for Volatility Estimation with GARCH Models