Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

I am working with $G_{n,p}$ random graphs. Let $p$ (the probability of an edge) be $c/n$, so $c$ is the expected number of neighbors for a vertex when $n$ is large.

Now let $N$ be the number of components with more edges than vertices in the graph. I am trying to find the largest $c$ such that $\mathbb{E}(N)$ tends to zero as $n$ grows large.

$\mathbb{E}(N) = \sum_{k=0}^n {n \choose k} \mathbb{E}(X_k)$,

where $X_k$ is the indicator that we have a component with $k$ vertices and $k+1$ edges or more.

$\mathbb{E}(X_k) = \mathbb{P}(\text{no outgoing edges}) \mathbb{P}(k+1 \text{ edges or more})$

$\mathbb{E}(X_k) = (1-p)^{k(n-k)} \mathbb{P}(\text{Binomial}({k \choose 2}, p) \ge k+1)$

Here I am stuck when trying to find an upper bound for this expression. The inequalities I've tried either diverge or cause trouble in the summation. Hoeffding's inequality looked promising but gave me $k^2$ in the exponent which gave me a hard time with the summation. Any tips?

share|improve this question
add comment

1 Answer 1

up vote 2 down vote accepted

The threshold for cycles in $G(n,p)$ is $p=1/n$. The same threshold should work in your case, so the solution is basically $c = 1$.

share|improve this answer
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.