Thread Subject: How to estimate ARMA-GARCH by QMLE

I want to estimate parameters in a ARMA-GARCH model by quasi-likelihood. But functions of the GARCH toolbox, such as garchfit, are all based on maximum likelihood.

How can I do the same thing under quasi-MLE? Do I have to programe it by myself? I just think someone else should have already solve it before. Many thx for your help!