Login

Welcome, Guest. Please login or register.

April 23, 2024, 04:08:35 pm

Author Topic: --  (Read 6511 times)  Share 

0 Members and 1 Guest are viewing this topic.

Planck's constant

  • Victorian
  • Forum Leader
  • ****
  • Posts: 748
  • Respect: +52
Re: FlorianK's OlympiadeMath-Thread
« Reply #15 on: June 06, 2013, 02:02:23 pm »
0
I'm working on that problem atm, I just got to the point that

ao=a1=1 and an+1=14an - an-1 - 4.



It seems to be true for n>1 when a0=1 and a1=1, but also when a0=1 and a1=9.
However it seems to also be true for any integers a1>a0, as long as n is large enough (17? - not sure I was experimenting with a spreadsheet)
So may be there is some hope for simple induction.

Nice problem Floriank

Alwin

  • Victorian
  • Forum Leader
  • ****
  • Posts: 838
  • Respect: +241
Re: FlorianK's OlympiadeMath-Thread
« Reply #16 on: June 06, 2013, 02:29:03 pm »
+1
Matrix Method (aka overly long and complicated method):



Apologies, did this on Word, since I prefer it over LaTex

It gives the general formula for an+1.... I think :P
2012:  Methods [48] Physics [49]
2013:  English [40] (oops) Chemistry [46] Spesh [42] Indo SL [34] Uni Maths: Melb UMEP [4.5] Monash MUEP [just for a bit of fun]
2014:  BAeroEng/BComm

A pessimist says a glass is half empty, an optimist says a glass is half full.
An engineer says the glass has a safety factor of 2.0

stolenclay

  • Victorian
  • Trendsetter
  • **
  • Posts: 133
  • AN, pls.
  • Respect: +78
Re: FlorianK's OlympiadeMath-Thread
« Reply #17 on: June 06, 2013, 10:19:48 pm »
+6
I'm not sure whether you've solved it, FlorianK, but with some of the progress you have made, I think I have it.
Spoiler
This solution isn't as elegant as could be... Anyway.
We have a recurrence relation with initial condition .

This recurrence is inhomogeneous (because of the ), but we can try homogenising it with a substitution , where is a constant.
We want the sequence to satisfy .

Now we have .
Solving our recurrence relation with initial condition , we obtain using characteristic equations and what-not (which I believe is quite standard in Olympiad Mathematics?).
Hence , which is consistent with Alvin's post above.
(Alvin has .)

Now, from your progress on the problem, you observed that the sequence seems to satisfy with .
If we define to actually be the recurrence satisfying with , then solving the recurrence relation gives .

Now the only thing left to do is to cross our fingers and hope we can prove that .


And finally, , so, indeed .

The only other thing to prove is that is actually an integer, which is easy, given that we defined it as the sequence with .
« Last Edit: June 10, 2013, 12:17:58 am by heaiyuo »
Thoughts on my journey through university
2014–2016 BCom (Actl), DipMathSc @ UoM
2017–2018 Master of Science (Mathematics and Statistics) @ UoM

kamil9876

  • Victorian
  • Part of the furniture
  • *****
  • Posts: 1943
  • Respect: +109
Re: FlorianK's OlympiadeMath-Thread
« Reply #18 on: June 06, 2013, 11:54:34 pm »
+7
Here is the solution I was alluding to earlier:

So let P(n) denote the statement:

: and

The base case P(1) can be easily verified from the initial conditions. Now let's prove the induction step. So suppose P(n) holds, then:



Which verifies the first part of P(n+1), now we need to verify the second part:



And hence the induction step is totally complete.
Voltaire: "There is an astonishing imagination even in the science of mathematics ... We repeat, there is far more imagination in the head of Archimedes than in that of Homer."

Planck's constant

  • Victorian
  • Forum Leader
  • ****
  • Posts: 748
  • Respect: +52
Re: FlorianK's OlympiadeMath-Thread
« Reply #19 on: June 07, 2013, 12:06:05 am »
0
Good work guys.
Kamil has now proven Frorian's 'observation' that

 sqrt(An+1) = 4sqrt(An) - sqrt(An-1)

And heaiyuo has done the rest

kamil9876

  • Victorian
  • Part of the furniture
  • *****
  • Posts: 1943
  • Respect: +109
Re: FlorianK's OlympiadeMath-Thread
« Reply #20 on: June 07, 2013, 12:13:29 am »
0
heaiyuo also proved this observation. We both proved the exact same thing just with different methods.
Voltaire: "There is an astonishing imagination even in the science of mathematics ... We repeat, there is far more imagination in the head of Archimedes than in that of Homer."

stolenclay

  • Victorian
  • Trendsetter
  • **
  • Posts: 133
  • AN, pls.
  • Respect: +78
Re: FlorianK's OlympiadeMath-Thread
« Reply #21 on: June 07, 2013, 12:47:11 am »
0
kamil9876's solution does this problem justice. Solving recurrences is mostly very mechanical, contradicting with the style of Olympiad Mathematics problems. I definitely like kamil9876's solution more than I like my own.

Of course, the major part in this was still FlorianK's motivation to consider the sequence , which, I (like to) believe, really needs a lot of mathematical instinct/intuition, as well as playing around a lot with the problem.
Thoughts on my journey through university
2014–2016 BCom (Actl), DipMathSc @ UoM
2017–2018 Master of Science (Mathematics and Statistics) @ UoM

FlorianK

  • Victorian
  • Forum Leader
  • ****
  • Posts: 928
  • Respect: +64
Re: FlorianK's OlympiadeMath-Thread
« Reply #22 on: June 07, 2013, 02:22:14 am »
0
well to unveil my thought process. I needed to find a sequence that gives an=bn²,  because well that is actually what we need to proof. So I looked at values of sqrt(an) which are, b needs to be 1 for n=0 and n=1 as well for obvious reasons.

n=0 a=1 b=1
n=1 a=1 b=1
n=2 a=9 sqrt(a)=3
n=3 a=121 sqrt(a)=11
n=4 a=1681 sqrt(a)=41

Then I found the progression in 1 1 3 11 41 :)

After I considered bn+1=4bn-bn-1 I just checked with the 2 numbers that followed.
Number sequences such as 1 1 3 11 41 come up in IQ tests quite often as well. I once had a phase where I was actually doing heaps of them for pleasure :p

Thx for the help guys :D, just one question



So let P(n) denote the statement:

: and
How did you get to the second equation?
« Last Edit: June 07, 2013, 02:34:17 am by FlorianK »

FlorianK

  • Victorian
  • Forum Leader
  • ****
  • Posts: 928
  • Respect: +64
Re: FlorianK's OlympiadeMath-Thread
« Reply #23 on: June 07, 2013, 07:28:23 am »
0
Hey, so this is for geometry.
So you have a triangle ABC and AB is the radius of a circle with A being the center of a circle. C is out side of the circle. CB cuts the circly at X, CA cuts the circly at Y. If you extend the line CA it'll cut the circle at Z

So for geometry questions with such a circle and triangle people use the relationship CX*CB=CY*CZ.
Does anybody have a link to a proof for this?

kamil9876

  • Victorian
  • Part of the furniture
  • *****
  • Posts: 1943
  • Respect: +109
Re: FlorianK's OlympiadeMath-Thread
« Reply #24 on: June 07, 2013, 11:05:41 am »
0


Thx for the help guys :D, just one question
How did you get to the second equation?

So we wanted to prove by induction. And so the inductive step would go something like this.




But now we want to this to be equal to and so I "solved" for in order to see what it should be. So then I decided to add that to my induction hypothesis.
Voltaire: "There is an astonishing imagination even in the science of mathematics ... We repeat, there is far more imagination in the head of Archimedes than in that of Homer."

stolenclay

  • Victorian
  • Trendsetter
  • **
  • Posts: 133
  • AN, pls.
  • Respect: +78
Re: FlorianK's OlympiadeMath-Thread
« Reply #25 on: June 07, 2013, 04:23:48 pm »
+1
Diagram
Proof
The quadrilateral is cyclic (and has to be non-degenerate as well, given that the "order" of all 6 points around has to be ).

A similar proof also works for being inside the circle, but certain letters have to be swapped, of course.
I think they call this power of a point in Olympiad Mathematics.
EDIT@below: Damn, I'd already reuploaded a smaller one. I guess I'll still resize this one though. Thanks for the tip!
« Last Edit: June 07, 2013, 04:40:39 pm by heaiyuo »
Thoughts on my journey through university
2014–2016 BCom (Actl), DipMathSc @ UoM
2017–2018 Master of Science (Mathematics and Statistics) @ UoM

b^3

  • Honorary Moderator
  • ATAR Notes Legend
  • *******
  • Posts: 3529
  • Overloading, just don't do it.
  • Respect: +631
  • School: Western Suburbs Area
  • School Grad Year: 2011
Re: FlorianK's OlympiadeMath-Thread
« Reply #26 on: June 07, 2013, 04:32:50 pm »
0
EDIT: Now I really wish I'd made the diagram smaller :(
Tip: Add "width=950" in the image tag, so that it should read:
Code: [Select]
[img width=950]http://i.imgur.com/bZPVR7s.png[/img]
EDIT: To compensate for the spoiler tag, make it 940 :)
2012-2016: Aerospace Engineering/Science (Double Major in Applied Mathematics - Monash Uni)
TI-NSPIRE GUIDES: METH, SPESH

Co-Authored AtarNotes' Maths Study Guides


I'm starting to get too old for this... May be on here or irc from time to time.