0

How can I show that $n^2 - 2n - 3$ is $\Omega(n^2)$?

1 Answers1

2

Since $2n+3$ is positive for any choice of $n$, $$n^2-2n-3\leq n^2$$ for any $n\geq 1$.

Now, $$-\frac 1 3 n^2\leq -3$$ and $$-\frac 1 3 n^2\leq -2n$$ when $n\geq 5$. Thus $$n^2-\frac 23n^2=\frac 13n^2\leq n^2-2n-3$$ when $n\geq 5$. We can then write $$\frac 13n^2\leq n^2-2n-3\leq n^2$$ for $n\geq 5$

Pedro
  • 122,002
  • How did you come up with the $-1/3$ coefficient? –  Jul 06 '13 at 01:56
  • Trial and error. – Pedro Jul 06 '13 at 01:57
  • You have to argue that is also true when $n\geq 2$. – Pedro Jul 06 '13 at 02:04
  • Sorry, that's what I meant. When $n \ge 2$, $-1 n^2 \le -3$. –  Jul 06 '13 at 02:06
  • So I took a break and came back to this problem. I came up with $C = -1$ and $k \ge 2$. Would those be appropriate witnesses to say that $n^2 - 2n - 3$ is $\Omega (n^2)$? –  Jul 06 '13 at 03:40
  • Update: My logic is: $n^2 \le n^2$ when $n>=1$ and $C=1$, $n^2 \le -2n$ when $n>=2$, and $C=-1$ and $n^2 \le -3$ when $n>=2$ and $C =-1$. Summing the coefficients results in $-1$ and the $max(k_1,k_2,k_3) = 2$. –  Jul 06 '13 at 03:42