Both F are the same, but your solve() is returning the wrong answer. The lambertw argument is of the form A*exp(-B) but your solve() is instead forming 1/A * exp(B) .
Which MATLAB versions are the two of you running? Is one of you running with the Maple based symbolic toolbox, and the other running with the MuPAD based symbolic toolbox?
Note: complex values are not inherently wrong solutions. The general solution is not restricted to LambertW branch 0 the way the two solutions here show.