Fuzzy greater than or equal to for floats.
Returns true if a is greater than or equal to b, allowing for epsilon error due to limitations in floating point accuracy.
a
b
epsilon
Does not handle overflow, underflow, infinity, or NaN.