In a paper I'm currently reading it gives alpha to be the following value.
$\alpha = \max_t \min_{t_j \in T_N} ||t-t_j||_2$
I am wondering what exactly this means? I have the following code:
alpha = max(tmax, min(tjmin, sqrt((t - tj) ** 2)))
where $tmax = \max(t)$, and $tjmin = \min(t)$
Is this the correct interpretation?
Read it inside out. First, suppose you had picked some $t$: then you could scan your available $t_j\in T_N$, and see which of these is closest to $t$. You could repeat that for a different $t$, and presumably get a different minimum distance in each instance. Among all these $t$ and their minimum distances to any of the $t_j$'s, there's at least one $t$ which will never be closer than any of the others, i.e. it has the largest minimum distance. That's $\alpha$.