Recursive Preferences in Dynare ++

Dear all,
I am trying to use Dynare++ to solve optimization problem with epstein -zin preferences (similar to Colacito Croce 2012) in Dynare ++.
How should I enter the objective function? It seems taking clue from Dynare++ user guide that once we enter the objective function and discount factor, dynare++ automatically considers it to be a infinite horizon problem with separable utility.
Because the example in user guide deals with separable utility function, I am not sure how can I modify that for recursive preferences.

Should I enter objective as U and define U in the model block?
Can anybody help?

See [Expected value of a power)

Thanks for the reply. I understand how to enter lagged expectation. My precise question was how to enter utility function in “Objective Block” .
Thanks in advance