I think the jury is still out on mathematical realism. Our current physics is almost certainly an imperfect description/model however.

Even if we'd know the perfect model for particle physics, we still can't apply math straight for simulation. We always prefer an approximation, all real world models are approximations.