Universal uniqueness criterion?

[bo198214]

Of course, this procedure might not help if the derivatives at the fixed points are positive reals (or are 0). In that case, near the fixed points, the expanded \( A_1 \) won't be any closer to \( A_2 \) than the original one was.

Right, this would make up an extra case; also it needs to be investigated whether this case allows an initial region at all. However for \( \exp_b \), \( b>e^{1/e} \) this is already satisfied.