“…Therefore, like the deterministic translations of LTL, it can be shown to construct doubly exponential sized automata even for very simple LTL fragments like those that contain ∨, ∧, and F. Thus, it does not achieve the optimal bounds for LTL\GU shown in [9]. However, one advantage of the construction in [21] is that it can be used in quantitative verification as well as qualitative verification of MDPs and has been implemented in [22]. Quantitative verification of MDPs can also be performed using nondeterministic automata that have the good-for-games (GFG) property [8,12], but translating a general NBA into a GFG automaton is known to result in an exponential blow-up.…”