simulate unstable ar model

Hello everyone!

For class I have to do a monte carlo simulation of an unstable AR process.

Unfortunately, when I program the model, it cannot estimate it since it is unstable:

alpha_1=0.7;
alpha_2 = 0.3; 
alpha = [alpha_1 alpha_2]; %AR parameter
cons = (1 - alpha_1 - alpha_2)*2; %constant of the AR process

model = arima('constant',cons,'AR',alpha, 'variance', 1);

I get the following error message after the last row:

"Nonseasonal autoregressive polynomial is unstable."

Anyone knows how to handle this?

Thanks in advance.