find fcns satisfying f(n) = Q(g(n)), h(n) = Q(t(n)), and

mathguy78

New member
Joined
Nov 28, 2006
Messages
9
I've been doing some work with algorithms and I've come across this problem and I just can't seem to get my head around it. Could anybody maybe steer me in the right direction with this?

Thanks for any help

Find functions satisfying
f (n) = Q(g(n))
h(n) = Q(t(n))
f (n)- h(n) =/ Q(g(n)- t(n))

Q = theta notation
=/ means not equal to, in the book, they are equals with a slash through them
 
What is the definition of \(\displaystyle \theta ( n )\, ?\)
 
pka said:
What is the definition of \(\displaystyle \theta ( n )\, ?\)


There isn't one, essentially, we just need to find a function that works for all 3 from what I understand.
 
The theta notation essentially means "about the same as."

It is considered the intersection of the Big-O and Big-Omega of a function.

For instance if you have a function F which is O(G) and Omega(G) then it is said to be Theta(G). F is O(G) means there is a real number \(\displaystyle \epsilon\) such that \(\displaystyle \epsilon\)G \(\displaystyle \ge\) F for some input size cutoff i.e. for large natural numbers n, \(\displaystyle \epsilon\)G performs worse than F.

Similarly, F is Omega(G) means there is a real number \(\displaystyle \delta\) such that \(\displaystyle \delta\)G \(\displaystyle \le\) F for some input size cutoff i.e. for large natural numbers n, \(\displaystyle \delta\)G performs better than F.

For a quick example, consider the functions f(n)=n<sup>2</sup> and g(n)=3n<sup>2</sup>.

f(n) is O(g(n)) because with \(\displaystyle \epsilon\)=1, \(\displaystyle f(n) \le\) \(\displaystyle \epsilon g(n)\). It is also Omega(g(n)) since \(\displaystyle \delta = \frac{1}{4}\) satsifies f(n) \(\displaystyle \ge \delta g(n)\). Therefore f(n) is Theta(g(n)).

There is also a calculus-based technique. take \(\displaystyle \L {lim}_{_{n \rightarrow \infty}} \frac{f(n)}{g(n)}\). If L is a constant number then f(n) is Theta(g(n). If it is zero then f(n) = O(g(n)). If it is infinity, then f(n)=Omega(g(n)).
 
Where else have you seen the big-O notated as theta?
 
(also, note that "=" here really means "has the property of being").

Consider:

\(\displaystyle f(n) = n^2 + n \\ h(n) = n^2 \\ g(n) = 2n^2 \\ t(n) = n^2\)

We have \(\displaystyle (f - h)(n) = n\) and \(\displaystyle (g-t)(n) = n^2\).

It is obvious that \(\displaystyle \L n \neq \Theta (n^2)\).
 
pka said:
Where else have you seen the big-O notated as theta?
Was this directed at me? If so, I'm not quite sure what you mean. The only time I've ever used Big-\(\displaystyle \L \Theta / O / \Omega\) was in my Algorithms class a year ago.
 
No. All I asked in the first place is for a definition.
I have never seen a big-\(\displaystyle \L \Theta\).
I do not think that it is standard.
 
pka said:
Where else have you seen the big-O notated as theta?


Yeah, the theta has to do with the algorithms


Thank you for the explanations, I appreciate your help.
 
daon said:
I believe it is a standard notation…[in] the book we used: Algorithm Design
I have no doubt that you are correct about Algorithm Design.
The big-O has been around mathematics for a century.
But I have just learned from Ken Rosen’s book that the big-\(\displaystyle \L \Theta\) was invented by Don Knuth about twenty-five years ago to overcome the ambiguities left by big-O. Not being in CS it is not standard to me.
 
Top