Gap-Trend model

Estimation ML/Bayesian estimation
4

Yes, for simulations you can impose any value you like. If you want to have a drift, you need a linear model and use the nocheck option of steady. See