Expected value of the number of vertices more than 2 edges apart in a random graph

59 Views Asked by At

We create a random graph with $n$ vertices in such a way that for each pair of vertices $\{i, j\}$, we connect them with an edge with probability $p$ (each time drawing independently). Let $X$ be the number of pairs of vertices $\{u, v\}$ such that the distance from $u$ to $v$ is greater than 2. For example, for the graph $G=(V, E), V=\{1,2,3,4,5\}, E=\{\{1,2\},\{2,3\},\{3,4\}\}$, we have $X=5$ (the pair $\{1,4\}$ and four pairs $\{i, 5\}$ for $i<5$).

a) Calculate $\mathbb{E}[X]$.

b) Calculate Var $X$.

I think that $$\mathbb{E}[X] = \binom{n}{2} - \mathbb{E}[Y_1] - \mathbb{E}[Y_2]$$

where $Y_1 =$ the number of pairs of vertices separated by a distance of $1$,
and $Y_2 =$ the number of pairs of vertices separated by a distance of $2$.

Then $$\mathbb{E}[Y_1] = \binom{n}{2} \cdot p$$

because each pair is connected with probability $p$, and

$$\mathbb{E}[Y_2] = \binom{n}{2} (1-p) (n-2) p^2$$

because each pair of vertices is not directly connected with probability $(1-p)$, and has $(n-2)$ vertices through which it can be indirectly connected with probability $p^2$ each.

However, for $n = 10$ and $p = \frac{1}{2}$, it turns out that $\mathbb{E}[Y_2] = \binom{10}{2}$, which is the number of all possible pairs, so I suspect that the same events are counted multiple times.

How can I find $\mathbb{E}[Y_2]$, and why is what I have now wrong (why can't we consider each pair of vertices independently thanks to linearity of expectation)?

1

There are 1 best solutions below

0
On BEST ANSWER

Your setup is generally good, and you’re right that you can consider each pair separately due to linearity of expectation.

But your expression for $\mathsf E\left[Y_2\right]$ is wrong. The $n-2$ events that the two vertices are connected via another vertex are independent, so the probability for none of them to occur is $\left(1-p^2\right)^{n-2}$, and thus

$$ \mathsf E\left[Y_2\right] = \binom n2(1-p)\left(1-\left(1-p^2\right)^{n-2}\right) \;. $$

For $n=10$ and $p=\frac12$, this is

\begin{eqnarray*} \mathsf E\left[Y_2\right] &=& \binom{10}2\left(1-\frac12\right)\left(1-\left(1-\left(\frac12\right)^2\right)^{10-2}\right) \\ &=& \frac{45}2\cdot\left(1-\left(\frac34\right)^8\right) \\ &=& \frac{2653875}{131072} \\[4pt] &\approx&20.25 \;. \end{eqnarray*}