Suppose we have an arbitrary sequence $$\{a_k\} = \{a_1, a_2, ..., a_k\} $$
and use it to a create a set as follows
$$A = (a_i+ a_j : a_i, a_j \in \{a_k\})$$
and we wanted to sum over all of the members of this set, would we denote it as $\sum A$ ?
Furthermore, if we wanted to create a new set $A^*$ which excludes terms $a_i + a_j$ if $i=j$, how could we write that more succinctly?
How about if we wanted to exclude terms $a_j + a_i$ aswell? i.e. $a_i + a_j$ in reverse order?
Ultimately, I'm looking for notation akin to summing over all "unique pairwise sums" of the members of a set (something like $\sum B$ where $b := a_i + a_j$ and $b \in B$) excluding the cases where $i=j$, and excluding all the "duplicate pairwise sums" of the form $a_j + a_i$ given that they're not "unique" ($a_j + a_i$ is given to be "equivalent" to $a_i + a_j$)
I'm assuming there is a more succinct way to write "the sum of all unique pair-wise sums of the members of a set", because I vaguely remember notation form probability theory to denote "all pairs from a set where order doesn't matter"... something to do with binomials if I remember correctly.
Apologies in advance for my drawn out description... Any better notation that captures what I'm trying to capture would be much appreciated.
Yes this is fine notation. You could also write $\sum_{a \in A} a$ or $\sum_{i = 1}^k a_i$.
$A^* = \{a_i + a_j : a_i, a_j \in \{a_k\} \text{ and } i \neq j\}$.
I'd denote this $\{a_i + a_j : a_i, a_j \in \{a_k\} \text{ and } i < j\}$. This will avoid this double counting of sums as well as the $a_i + a_i$.
In the case here, I'd write $\sum_{i < j} a_i + a_j$. The same works for any set indexed by something totally ordered.