((Let's review, then, and the Toaster totally-not-arbitrary rating of ability to recreate.
Mission 1 [Crystal Alien Ark]: The internal structure was mostly mapped, IIRC. All it needs is AI for the other teams and the roving bots. Rating: A
Mission 2 [Erik and Friends]: Hmm... I think there's enough data to recreate what is going on with the illusions, or at least to make logical leaps there. IIRC the temple itself wasn't fully explored, plus there were zero attempts to negotiate with Erik, so it's unclear what his motives were. Rating: B
Mission 3 [Aliens in my Ear]: Other than what would have happened if someone tried to enter the pool, I think this one's pretty much gold. Rating: A
Mission 4 [Damn Dams]: Despite the Dungeon Bypass executed here, the potential for randomization of the resistance faced actually makes this probably the best of the bunch for reproduction, given the high quality map data we had going in. Rating: A+
Mission 5 [Serial Killer Spring Break]: While randomizing the placement of the killer makes things interesting, the unexplored town itself and the zero insight to his motives makes this one tricky. Rating: C
Mission 6 [The Long Dark]: There's a whole lot of factors here that are totally unknown: the Lost Citadel environs, how a lot of things would have interacted, the Wyrm and Lighteaters... yeah, no. Rating: D
Mission 7 [Roboland]: While there was a lot unexplored, the fact that the team pretty much completely beat this mission and got back all the data may make this one possible. Provisional based on how much the recovered data can make its way to VR. Rating: B+ or C-, depending.
Mission 8 [Sharkmist Attack!]: We were pretty thorough in exploring this one. The only question mark is what was down the final tunnel. The size of the team required may hinder this one, though. Rating: B
Conclusion: Totally feasible. Even just the simulation of Mission Four would be a great training ground. You could even simulate just one half of the mission and let AI handle the other half.))