Theres always allowing the ability to try importing designs, but that raises its own issues. Mostly trying to keepsome form of limits (say, by having a set of rolls for getting authorization, permission from the company/nation that makes it, availability, licensing) and then it depends on the players to even remember they can do it.
Limiting the number of design actions has merit, but then its nearly impossible to keep limits of them that make sense, arent over or under powered, and dont swamp the GM.
Now, combining numbers 1 and 3, to let there be pools of workers who can be set to work on a number of projects, is the worst in practice, and the best in theory, atleast mostly. However, if you want there to be any form of progression in the work force at all, the book keeping gets oppressive quickly, to the point where it becomes work, hard, aggravating work, to keep track without some form of program to help with it.
One of the best weapon design game I have seen was a mixture of 1, 3, and 4, wherein the players represented the management of one design lab in a nation, focusing mostly on a few aspects of the military, although they quickly started trying to do everything and the other groups became mostly just story aspects. Plus, the GM then has to select everything else that gets designed, which puts a fair amount of bias on some designs.
There was also one where each player was a corporation, making designs costed money, and then the designs had to be produced, but, well, that dident last long. A similar system, wherein the group is, say, an independent factory of some flavor, has some merit, but then its somewhat stifling and the bookkeeping, my god, the bookkeeping...