Good question! I'll try to simplify the situation for you. Suppose, $A = \{1,2,3,4\}$. Also suppose there is a function $f: A\to\mathbb R$, such that $$f(1) = 0, f(2) = 3, f(3) = -2, f(4) = 1$$
I shall explain what $$\max_{a\in A} f(a)$$ means! Well, it essentially asks you to iterate over set $A$, compute the value $f(a)$, and return back the maximum value you see in the process. In our case, $\max_{a\in A} f(a) = 3$. Another useful notion is that of $\text{argmax}$. $$\operatorname{argmax}\limits_{a\in A} f(a) = 2$$
$\text{argmax}_{a\in A}f(a)$ returns the element(s) from set $A$, for which $f(a) $ is maximum. On the other hand, as you saw, $\max_{a\in A}$ returns the maximum value itself! Hope that clarifies the notation.
P.S. Isn't that a Bellman equation in Reinforcement Learning? Super cool, have fun!