AMC Registration in 24 days
Mon December 15, 2008

Suffolk County Math Tournament in 49 days
Fri January 9, 2009

PUMaC in 71 days
Sat January 31, 2009

AMC 12A in 81 days
Tue February 10, 2009

HMMT in 92 days
Sat February 21, 2009

World Math Day in 103 days
Wed March 4, 2009

AIME I in 129 days
Mon March 30, 2009

PLUS! magazine deadline in 130 days
Tue March 31, 2009

USAMO in 158 days
Tue April 28, 2009

ARML in 189 days
Fri May 29, 2009

21st Birthday in 660 days
Sun September 12, 2010

Turning 40 in 7600 days
Wed September 12, 2029

tags in Topics

more tags

Feeds

Navigation

Links

If you liked this post, buy me a beer... I mean cookie.

Towers of Powers

Posted June 2nd, 2008 by Mgccl

in

ARML
Math

I ARML 2008ed! It was great, definitely doing it next year.
There was an ARML Lecture at 8pm on the day before the ARML. Sadly I was feeding at that time, with 3 other students from my school and a Intel Talent Search Finalist who also got admitted into MIT and Harvard(She chose Harvard over MIT...) and had the highest amount of individual question right on the ARML meet in our Suffolk County Allstar team.
Our team t-shirt was: "We have imaginary friends." and a face created with a unit circle, a segment of a hyperbola and imaginary numbers. Yeah, that's the best non-perverted math t-shirt we can think of...
I am god at math jokes... most math jokes are only funny when they are perverted... I actually collect math jokes as hobby... If you need any you can ask for some
For the first time, I slept inside a university, had Subway, had Arby's, had something that I believe it's omelet.
I gained a huge amount of confidence. 4 months with 14 hours per day on math training is about enough time to make to USAMO. Yeah!

The lecture was on two problems:

1. Given , does the sequence , , $n^{n^n}$ ... become constant when consider in the system $\mathbb{Z}/k\mathbb{Z}$ of integers modulo

2. Given , does the sequence , ,... periodic mod ? If so, when does it repeat?

Since I'm not in the lecture, I can only try to figure it out myself.
I think solved the first one, haven't start the second one.
Please anyone... check if my solution is correct or not.

Proof:
For the number in the sequence mod k eventually become a constant, then there must be a number so that $x^n \equiv x\ (mod\ k)$ .
so the next integer in the sequence mod k is.
$\begin{eqnarray} x^n^n &=& \underbrace{(((x^n)^n)^n....)}_{n n's}\\ \underbrace{(((x^n)^n)^n....)}_{n n's} &\equiv& x\ (mod\ k) \end{eqnarray}$
and so on.

After finding the property of the number, now I have to find if that number ever exist in the sequence.
Answer is yes.
For any $a^i \equiv b\ (mod\ k)$ , keep increasing i will eventually create a set of repeating numbers, and the size of the set is smaller than k.
With this information, must exist
Suppose $a_i = b^i mod\ k$ .
a_i have a period of , all the possible numbers are in the ordered sequence .
If d=n , trivial case, will be accessed the at the first try, and it's 0.
If $d\not{=}n$
$a_n, a_{n^n}...$ will go though all the number in . Until eventually one land on the .

I'm too sleepy, can't form a great argument.

Bookmark/Search this post with:

Question 1

On Jun/02/2008 15:56:01 Simon (not verified) says:

I'm not convinced by your solution, sorry

reply

I actually need paper and

On Jun/02/2008 17:58:00 Mgccl says:

I actually need paper and graph to show it... because I don't know how to actually present it. I will post a newer version as soon as I finish my school ends