Rules of Double Sums

Solution 1:

The picture for rule 1 looks like this:

$$ \begin{array}{c|ccccc} & x_1 & x_2 & x_3 & x_4 & x_5 \\\hline y_1 & x_1y_1 & x_2y_1 & x_3y_1 & x_4y_1 & x_5y_1 \\ y_2 & x_1y_2 & x_2y_2 & x_3y_2 & x_4y_2 & x_5y_2 \\ y_3 & x_1y_3 & x_2y_3 & x_3y_3 & x_4y_3 & x_5y_3 \end{array} $$

That is, to find $(x_1 + x_2 + x_3 + x_4 + x_5)(y_1 + y_2 + y_3)$ we can look at all possible products $x_iy_j$ and sum them. The products can be arranged in a rectangle as above.

When the $y$'s equal the $x$'s we get

$$ \begin{array}{c|ccccc} & x_1 & x_2 & x_3 & x_4 & x_5 \\\hline x_1 & x_1^2 & \color{red}{x_2x_1} & \color{red}{x_3x_1} & \color{red}{x_4x_1} & \color{red}{x_5x_1} \\ x_2 & \color{purple}{x_1x_2} & x_2^2 & \color{red}{x_3x_2} & \color{red}{x_4x_2} & \color{red}{x_5x_2} \\ x_3 & \color{purple}{x_1x_3} & \color{purple}{x_2x_3} & x_3^2 & \color{red}{x_4x_3} & \color{red}{x_5x_3} \\ x_4 & \color{purple}{x_1x_4} & \color{purple}{x_2x_4} & \color{purple}{x_3x_4} & x_4^2 & \color{red}{x_5x_4} \\ x_5 & \color{purple}{x_1x_5} & \color{purple}{x_2x_5} & \color{purple}{x_3x_5} & \color{purple}{x_4x_5} & x_5^2 \\ \end{array} $$

So what we see is that

$$ \color{red}{\mathsf{RED}} = \color{purple}{\mathsf{PURPLE}} \text{ and } \color{red}{\mathsf{RED}} + \color{purple}{\mathsf{PURPLE}} + \mathsf{BLACK} = \left( \sum_{i = 1}^m x_i \right)^2. $$

Hence

$$ 2 \color{purple}{\mathsf{PURPLE}} + \mathsf{BLACK} = \left( \sum_{i = 1}^m x_i \right)^2 \tag{1} $$

where

$$ \color{purple}{\mathsf{PURPLE}} = \sum_{i < j} x_ix_j \text{ and } \mathsf{BLACK} = \sum_{i = 1}^m x_i^2. \tag{2}$$

$(1)$ and $(2)$ imply

$$ \sum_{i < j} x_ix_j = \frac12 \left[ \left( \sum_{i = 1}^m x_i \right)^2 - \left( \sum_{i = 1}^m x_i^2 \right) \right]. $$

This shows up whenever you want to evaluate a sum in two indices where one index is $<$ the other. This shows up for instance in probability and combintorics.

Solution 2:

Explanation to (3): Given relation is

\begin{align} \left(\sum_{i = 1}^{n} x_{i} \right)^{2} & = \sum_{i = 1}^{n} x_{i} \sum_{j = 1}^{j} x_{j} \tag{1} \\ & = \sum_{i = 1}^{n} \sum_{j = 1}^{n}x_{i}x_{j} \tag{2} \\ & = \sum_{i = 1}^{n}x_{i}^{2} + \sum_{i = 1}^{n}\sum_{\substack{j = 1 \\ j \neq i}}^{n} x_{i}x_{j} \tag{3} \end{align}.

To understand this consider a simple (still general) example of $n = 3$. So you have

\begin{align} \left(\sum_{i = 1}^{n = 3} x_{i} \right)^{2} & = \left( x_{1} + x_{2} + x_{3} \right)^{2} = x_{1}^{2} + x_{2}^{2} + x_{3}^{2} + 2 x_{1}x_{2} + 2x_{2}x_{3} + 2x_{3}x_{1} \\ & = \underbrace{x_{1}^{2} + x_{2}^{2} + x_{3}^{2}}_{\text{squared terms}} + \underbrace{x_{1}x_{2} + x_{1}x_{2} + x_{2}x_{3} + x_{2}x_{3} + x_{3}x_{1} + x_{3}x_{1}}_{\text{cross-product terms}} \tag{4} \end{align}

Now consider the following

\begin{align} \sum_{i = 1}^{n = 3} x_{i} \sum_{j = 1}^{j = 3} x_{j} & = (x_{1} + x_{2} + x_{3}) (x_{1} + x_{2} + x_{3}) \\ & = x_{1}^{2} + x_{1}x_{2} + x_{1}x_{3} + x_{2}x_{1} + x_{2}^{2} + x_{2}x_{3} + x_{3}x_{1} + x_{3}x_{2} + x_{3}^{2} \\ & = \underbrace{x_{1}^{2} + x_{2}^{2} + x_{3}^{2}}_{\text{squared terms}} + \underbrace{x_{1}x_{2} + x_{1}x_{2} + x_{2}x_{3} + x_{2}x_{3} + x_{3}x_{1} + x_{3}x_{1}}_{\text{cross-product terms}} \tag{5} \end{align}

Looking at (4) and (5) you can see that equality in (1) is true and the proof can be extended to any arbitrary number $n$.

Similarly,

\begin{align} & \sum_{i = 1}^{n = 3} \sum_{j = 1}^{n = 3}x_{i}x_{j} \\ = & \sum_{i = 1}^{n = 3} \left[ x_{i} x_{1} + x_{i} x_{2} + x_{i}x_{3} \right] \\ = & \left[ x_{1} x_{1} + x_{1} x_{2} + x_{1}x_{3} \right] + \left[ x_{2} x_{1} + x_{2} x_{2} + x_{2}x_{3} \right] + \left[ x_{3} x_{1} + x_{3} x_{2} + x_{3}x_{3} \right] \\ = & \underbrace{x_{1}^{2} + x_{2}^{2} + x_{3}^{2}}_{\text{squared terms}} + \underbrace{x_{1}x_{2} + x_{1}x_{2} + x_{2}x_{3} + x_{2}x_{3} + x_{3}x_{1} + x_{3}x_{1}}_{\text{croos-product terms}} \tag{6} \end{align}

Hence (6) and (4) explains the equality in (2). Now equality in (3) is merely the separation of the squared terms and the cross product terms.

\begin{align} & \sum_{i = 1}^{n = 3}x_{i}^{2} + \sum_{i = 1}^{n = 3}\sum_{\substack{j = 1 \\ j \neq i}}^{n} x_{i}x_{j} \\ = & \underbrace{(x_{1}^{2} + x_{2}^{2} + x_{3}^{2})}_{\text{squared terms}} + \underbrace{\left[ \underbrace{x_{1} (x_{2} + x_{3})}_{i = 1, j \neq 1} + \underbrace{x_{2}(x_{1} + x_{3})}_{i = 2, j \neq 2} + \underbrace{x_{3}(x_{1} + x_{2})}_{{i = 3, j \neq 3}}\right]}_{\text{cross-product terms}} \tag{7} \end{align}

Hence (7) and (4) explains the equality in (3) and this completes the explanation.