I'll bet it would be possible, with a good bit of work, to determine this using computer simulation. Trying to figure out the question "How many" is a lot harder than "How long can you run?", "How much strength does it take to knock out a 5 year old?", "How many five year olds can attack someone at once?", "How hard can an average 5 year old kick/bite/etc?". It should be much easier for people to agree with a relatively close number on those questions. Use those answers to have factors like endurance, endurance regeneration, strength, etc... plug them in to some type of computer simulation and ta da, problem solved.
Would there be any interest in a project like this?
I spend my entire day developing, running, and analyzing the results of computer simulations of many different structures, both static and dynamic. It is my professional opinion that it would be damn near impossible to get a realistic and accurate simulation like this, and if you could do it, it would take an investment on the order of hundreds of thousands of dollars (at least if you hired out the equivalent of me/my company to do the work). I really think conjecture is a more accurate and cheaper way to assess this situation.
To the recent poster who suggested an army of 5-year-old-size 2+2ers, I would not want to take on that challenge. The number for this challenge would probably be in the single digits - low double digits at best.