How to handle constraints?

Hi all,
I'm a beginner in the world of GA, so I beg your pardon if my terminology will be not exact.
I would have your opinion and suggestions about this problem: I have a binary genome of lenght M (that represents the coefficients of a known, complex function); I would set only N < M alleles to 1 using GA in order to maximize (or minimize) the function, so the problem is how to handle the contraint about the number of 1s in the genome.
For what I have studied, I know that I can use penalty functions (so that, penalize the fitness of "unfeasible" solutions) but I would know if there are other strategies to handle this type of problems.

Thanks in advance.

Regards
Libra
.