Is it always the case with genetic algorithms for optimization that equality constraints do not hold exactly?

Hi everyone, I programmed an optimization problem using genetic algorithms on MATLAB. I am modelling the weights of a portfolio of stocks, and as a linear constraint I have that all weights sum up to 100%. However, the weights that come out from the solutions sum up to something like 96% or 102%, etc. Is this due to the nature of genetic algorithms or is it that I have a mistake with the programming? Thanks ;)