GARCH forecasting model=ugarchspec(variance.model=list(model="sGARCH",garchOrder=c(1,1)),mean.model=list(armaOrder=c(1,1),include.mean=TRUE),distribution.model='sstd',start.pars = list(?), fixed.pars = list(?))
modelfit=ugarchfit(spec=model,data=r)
modelfor=ugarchforecast(modelfit,data=NULL,n.ahead=10,n.roll=0,out.sample=0)

I am forecasting future returns and vol for a fund for which I have only have 150 historical returns of monthly data. I was confused as what to put in the fixed and start parameters in ugarchspec function. Also if I want to do Monte Carlo simulation and forecast, is it possible? I found a function ugarchsim but didn't get how it works. 
 A: Firstly, the parameters are optional. Documentation to ugarchspec says that start parameters are not required unless the algorithm fail converging. Nevertheless, they may be set e.g. as start.pars=list(beta1=0.9, alpha1=0.05), which are all sensible values for "volatility with low volatility".
Once you fit the model and get the parameters, the forecasting is done in a way that you need to input the previous-day returns and volatilities. It is done through ugarchforecast, but you need to adjust your code for it to be useful.
model=ugarchspec(variance.model=list(model="sGARCH",garchOrder=c(1,1)),mean.model=list(armaOrder=c(1,1),include.mean=TRUE),distribution.model='norm', start.pars = list(beta1=0.93, omega=0.0000004, alpha=0.05))

modelfit=ugarchfit(spec=model,data=r, solver="hybrid", out.sample=20)

modelforecast<-ugarchforecast(modelfit, data = NULL, n.ahead = 1, n.roll = 5, external.forecasts = list(mregfor = NULL, vregfor = NULL))

The default rule for ugarchforecast is that only the first period prediction is based on the previous values, while for longer forecast the unconditional variance is calculated. So if you want to feed previous values into the forecast (which you probably want), then 
out.sample must be specified as an argument of ugarchfit. In ugarchforecast, the n.roll must be also specified, and it cannot be longer than the out.sample. Documentation for rugarch discusses this in detail. 
