The definition of simulation that is usually used includes the fact that simulation involves a certain model (usually, mathematical). The models might vary in their granularity and definition of what is considered key characteristics of the object/system/process they are intended to represent.
Then, during the simulation one is going to achieve (provided perfect implementation) a result that will be a subject to certain approximations being made (modeling error, measurement error, numerical method error, etc.) depending on the chosen model and simulation methodology.
Thus, depending on what you chose to be a "good enough" model of the universe, the answer to your question will be different.
For example, if you insist that the model of the universe must include the fact of the particular universe simulation (the one you are doing) happening in the future, you would arrive in a circular argument fallacy, and such simulations are not possible -> you would have to know the simulation result in advance.
I would also suggest a quick look at the Wikipedia article on simulation and Wikipedia article on mathematical models. It's worth to note that in order for a simulation to be perfect, the used model has to be perfect. If that seems to be your key question, you would have a more fruitful discussion on MathOverflow and Philosophy SE, while Physics might not be the best fit.