Consider replacing the current CE-delta measure for hop/route-change (based on absolute difference of previous and current Cross-entropy vaules) with Kullback–Leibler divergence (https://en.wikipedia.org/wiki/Kullback%E2%80%93Leibler_divergence).
They are quite similar but KL-divergence is perhaps better...?
Consider replacing the current CE-delta measure for hop/route-change (based on absolute difference of previous and current Cross-entropy vaules) with Kullback–Leibler divergence (https://en.wikipedia.org/wiki/Kullback%E2%80%93Leibler_divergence).
They are quite similar but KL-divergence is perhaps better...?