I am reading the basic properties of Concave Minimization from Global Optimization: Deterministic Approaches by Horst and Tuy and I came across such a notation on p. 10:
$$0 < \epsilon < \text{min}\{-a_i, b_i : i = 1, \dots, n\}$$
I am having a hard time interpreting the RHS of the inequality above. If $\epsilon$ was a vector, I would speculate that this notation would roughly mean: "choose the minimum of $-a_i$ and $b_i$ for $i$th element of the $\epsilon$. But I believe $\epsilon$ is a scalar here. So, what does this "minimum" notation really convey?