The only way to make the policy workable is to take the loved ones as well. If you're going to build a long term self contained colony you also have to think about the psychological wellbeing of the personnel, that means relationships, it means friendships, it means not taking people who are going to really get other people's backs up.
You can't ask people to leave their lovers and children behind, you have to take them as well. For each person you want to take, you need to take the entire immediate family. If this hasn't been considered then whoever ran the project is an idiot and deserves to fail epically.
You can leave the pseudo celebs behind, especially when you're talking about leaving all the ordinary people anyway. Much more important to take Randall Munroe.
Apart from the useless pseudo celebrities, you're unlikely to get much dead weight. People in the intellectual upper 10% don't tend to hang around with people from the bottom 10%. Whoever they ask to bring is probably going to be useful in the long term, if not the short term. And even then, any colony requires rubbish collection and cleaners, hairdressers and telephone sanitisers. There's work to be done that doesn't require an IQ of over 130.
Let's consider team selection, since that's the game we're playing here. In my spare time I do team sports (I know that's terribly bad form for a geek but that's how I roll) and as often as not, I'm the team captain so I have to choose my team.
The best player. He knows he's the best player, he likes to make sure everyone else knows he's the best player, you've all met him, he has the looks and the ego to match, but he's not the best player because his ego gets in the way of him being a team player. So he's out and still doesn't understand why.
The guy who turns up to everything and never seems to get any better. He might be in or out, depending how much other people like him, in my case he's a real charmer and usually goes. I took him to the last tournament and he played an absolute stormer, it happens.
The really good player with a +1 who will only play if she can as well, but she's not as good. This is where I'm leading to with this example because it's the one that matches your situation. This player isn't one person but two, sometimes two who work really well together even if the individual skills aren't up to the spec of no 1, the average and team benefit is greater than taking any two others.
You're not considering the skills of the individual but the combined skills and combined resource drain. The guy with a wife and a dozen kids, is he really worth 14 places when compared to a guy with 2 kids who's nearly as good? Probably not, if the second guy's wife has a similar or complementary skillset, definitely not. Every single position and person has to be considered like this. Not as a single person, but a combined skillset and cost of the total group. Is Steve worth Marco? That's up to how much you want Steve, but at one in a million, I suggest taking one of the other 300 candidates in the country.
The answer in most cases for special requirements versus someone else who's nearly as good but lower cost, probably not worth it.
Ultimately what this comes down to is that, as with any team selection, there are people who, on paper, you'd like to take, but you'll have to leave behind, and people who wouldn't be your first choice, but ultimately you will be taking them. People aren't numbers, you can't treat them that way, and the tighter your resources the more you'll be spending on reviewing each person as an individual, not under a general rule.