How to Write Multiple Sums with Complicated Indices

When several summation limits depend on one another, the nested notation can become difficult to scan. A useful alternative is to describe all allowed index tuples as a set, then sum over that set. This separates the question “which indices are included?” from “what is being added?”

There is no single best notation for every situation. Use a set-based sum when the constraints are the main source of clutter; keep the iterated sum when the order or structure of the summation is useful.

Replace complicated limits with an index set

Suppose a sum uses integer indices $i$, $j$, and $k$, with bounds that depend on earlier indices. Collect the restrictions into a set $I$ of allowed triples. Then write

$$\sum_{(i,j,k)\in I} a_{ijk}.$$

The notation says to include one term $a_{ijk}$ for every triple $(i,j,k)$ in $I$. If the set is not already clear, define it explicitly, for example:

$$I=\{(i,j,k)\in\mathbb{Z}^3: 1\le i<j\le n,\ 1\le k\le i+j\}.$$

This is equivalent to writing all three sums with their limits. It is often easier to check the conditions in the set than to follow a long chain of nested bounds.

A fully worked example

Consider the finite sum

$$S_n=\sum_{i=1}^{n-1}\sum_{j=i+1}^{n}\sum_{k=1}^{i+j}x_{ijk}.$$

The allowed indices satisfy $1\le i<j\le n$ and $1\le k\le i+j$. Define

$$I_n=\{(i,j,k)\in\mathbb{Z}^3:1\le i<j\le n,\ 1\le k\le i+j\}.$$

The same sum can now be written as

$$S_n=\sum_{(i,j,k)\in I_n}x_{ijk}.$$

For a concrete check, take $n=3$ and let $x_{ijk}=i+j+k$. The allowed pairs $(i,j)$ are $(1,2)$, $(1,3)$, and $(2,3)$. Their respective $k$ ranges are $1$ through $3$, $1$ through $4$, and $1$ through $5$. Thus

$$\begin{aligned} S_3 &=(4+5+6)+(5+6+7+8)+(6+7+8+9+10)\\ &=15+26+40\\ &=81. \end{aligned}$$

The set-based notation gives exactly the same 12 terms. It does not change the sum; it only packages the index restrictions differently.

Other useful forms

For a one-off expression, you can put the conditions directly beneath the summation symbol:

$$\sum_{\substack{1\le i<j\le n\\1\le k\le i+j}}x_{ijk}.$$

This is compact, but it can become cramped if the conditions are long. If the same collection of indices appears repeatedly, naming it—such as $I_n$—usually improves readability.

If the indices range independently over a Cartesian product, ordinary nested limits are often clearest. For example, when $1\le i\le m$ and $1\le j\le n$ independently, write $\sum_{i=1}^{m}\sum_{j=1}^{n}a_{ij}$, or equivalently sum over $(i,j)\in\{1,\ldots,m\}\times\{1,\ldots,n\}$. The set form is more useful when the allowed pairs are restricted, such as $i<j$.

When to keep the nested sum

Nested sums show dependencies directly. In the example, the upper limit for $k$ is determined by $i+j$, so the iterated notation makes that relationship visible. It can also be the better choice when you plan to evaluate the sum one index at a time or use an argument that depends on the summation order.

For finite sums, changing the order of summation does not change the result, provided you include the same terms. For infinite sums, changing or suppressing the order can change the meaning or value when the series is not absolutely convergent. In that setting, keep the order explicit unless a theorem or a stated convention justifies rearranging the terms.

Common mistakes

• Leaving the domain implicit. State that indices are integers or natural numbers when it is not clear from context. A condition such as $i<j$ alone does not specify the allowed values.
• Changing strict inequalities accidentally. The constraints $i<j$ and $i\le j$ include different index tuples. Check endpoints when translating limits into a set.
• Forgetting empty ranges. A nested sum may have no terms for some outer-index values. The standard convention for a finite empty sum is $0$; the set-based version handles this naturally because it contributes no elements.
• Treating an infinite sum as order-free. A compact sum over tuples should not obscure a meaningful iterated order for a non-absolutely convergent series.
• Using an unexplained shorthand. If you write $\sum_{\alpha\in I}a_\alpha$, define what an index $\alpha$ contains and what belongs to $I$ the first time it appears.