 mersenneforum.org Triangle puzzle
 Register FAQ Search Today's Posts Mark Forums Read  2006-05-17, 17:24 #1 dsouza123   Sep 2002 2·331 Posts Triangle puzzle This puzzle starts with a equilateral triangle with each side 1 unit in length. The midpoints of each side are connected forming more triangles. At each step, what is the count of same size triangles and the total length of all segments. List until the number of triangles is 1024 congruent equilateral triangles. The first two iterations: Count 1 Length 3 Each side 1 unit Count 4 Length 4.5 Each side 1/2 a unit.   2006-05-17, 23:24 #2 dsouza123   Sep 2002 10100101102 Posts The attached image file shows the first two iterations of the triangle. Attached Thumbnails   2006-05-18, 07:19 #3 Kees   Dec 2005 3048 Posts # triangles = 1024 # unit length of a triangle = 1/63 # sides of triangles (ignoring multiplicity) = 3*1024=3072 # length of sides (ignoring multiplicity)= 3072/63=1024/21 # length of sides of original triangle = 3 # multiplicity part of sides = 1024/21 - 3 = 961/21 (multiplicity =2) # real length of sides = 961/42+3= 1087/3= 362 + 2 / 3    2006-05-18, 10:45 #4 Kees   Dec 2005 22×72 Posts    never ever anwser a quick calculation 5 minutes before an important meeting. Ugh, terrible error at the end 961/42+3=1087/42=25+37/42 sorry    2006-05-19, 16:24   #5
mfgoode
Bronze Medalist

Jan 2004
Mumbai,India

22·33·19 Posts Symmetry

Quote:
 Originally Posted by Kees   never ever anwser a quick calculation 5 minutes before an important meeting. Ugh, terrible error at the end 961/42+3=1087/42=25+37/42 sorry  Forgive me pointing this out but a symmetrical problem requires a symmetrical answer.
On inspection of your method and answer I strongly believe that the more likely answer is 25.5 units.
It might be something else as I have not taken a shot at it but it seems more plausible.
Reminds me of the anecdote of Hardy perusing thru the papers of Ramanujan which were submitted from India and they had not met before.
He said to the effect that the identities must be true as who else could think of such symmetry and be wrong ?
Mally    2006-05-20, 06:51 #6 Kees   Dec 2005 22·72 Posts Well, cross checking in that case: we have 1024 triangles. If we count them from the base up there are 63+61+59+...+3+1=1024 of them So the base (ah, getting your point ) is occupied by 32+31 triangles, so unit length of triangle is 1/32 (which makes more sense indeed). So total unit length (without taking multiplicity in consideration) is 1024 (triangles) * 3 (sides) * 1/32 (unit length) = 96 The sides that are on the edge are not counted double and their length totals 3 (length of sides of original triangle). The remainder 93=96-3 are counted twice because they all form the border between two triangles. So divide by 2 This gives 46.5 Adding 3 gives 99/2    2006-05-20, 10:06 #7 drew   Jun 2005 2×191 Posts Each iteration will quadruple the number of triangles and halve the length of the segments. So, assuming n=0 is the iteration representing a single triangle with 1 unit sides # of triangles = 4n length of segments = (1/2)n Each iteration adds 3 segments to each of the previous iteration's triangles. If Ln represents the total length of the segments Ln = Ln-1 + 3 * 4(n-1) * (1/2)n For 1024 triangles, 4n = 1024. n=5...so we're at iteration 5. The length of each segment is 1/25 = 1/32. Now we just need to find the total length of the sides. We can do this iteratively. L1 = 4.5 L2 = 7.5 L3 = 13.5 L4 = 25.5 L5 = 49.5 Drew Last fiddled with by drew on 2006-05-20 at 10:07   2006-05-20, 10:27 #8 drew   Jun 2005 2×191 Posts If we don't want to iterate, we can continue with the following: Ln = Ln-1 + 3 * 4n-1 * (1/2)n = Ln-1 + 3/4 * 2n Ln = 3/4 (2n + 2n-1 + 2n-2 ... + 20) + C = 3/4 * (2n+1 - 1) + C Solve for C (L0 = 3) 3 = 3/4 + C C = 2.25 Ln = 3/4 * (2n+1 - 1) + 2.25 We can simplify this a bit further: Ln = 3/4 * 2n+1 + 3/2 = 3/2 * (2n + 1) Last fiddled with by drew on 2006-05-20 at 10:28   2006-05-21, 16:05   #9
mfgoode
Bronze Medalist

Jan 2004
Mumbai,India

22×33×19 Posts Multiplicity and Triangles

Quote:
 Originally Posted by drew Each iteration will quadruple the number of triangles and halve the length of the segments. So, assuming n=0 is the iteration representing a single triangle with 1 unit sides # of triangles = 4n length of segments = (1/2)n Each iteration adds 3 segments to each of the previous iteration's triangles. If Ln represents the total length of the segments Ln = Ln-1 + 3 * 4(n-1) * (1/2)n For 1024 triangles, 4n = 1024. n=5...so we're at iteration 5. The length of each segment is 1/25 = 1/32. Now we just need to find the total length of the sides. We can do this iteratively. L1 = 4.5 L2 = 7.5 L3 = 13.5 L4 = 25.5 L5 = 49.5 Drew Hats off to you Drew! this is really beautiful mathematics- to condense complicated thought into a minimum of symbols- the sheer economy of words- is mind boggling!
In this problem the difference in the lengths of the various inputs follows a simple law - the G.P. 1.5, 3, 6, 12, 24 corresponding to lengths [3,] 4.5,7.5,
13.5, 25.5 ,49.5 The next will be 49.5 + 48 and one can go on and on by just doubling the difference each time. Thanks to you Kees to help me convince myself that Drews iteration stands the test.
But your multiplicity (Never heard the term before, we called it duplication )
opened a can of worms for me and these last few days I spent staring at an equilateral triangle of 4^3 or 64 triangles trying to find some pattern with the previous ones.

But in all fairness to you your last post hit the nail on the head as I wanted to prove to myself the correctness of Drew's solution.

Of course the sheer symmetry of his answers no doubt convinced me to the hilt So we can put paid to this problem.

A word for Desousa 123. I have about 150 books on math recreations on my shelves but besides the 'circle dot' problem we discussed when I first joined, this is by far the second best I have come across. I am also curious to know as compiler what your solution is to this problem.

If you have a site for these problems I would love to have a go.
Thank you one and all.

Mally    2006-05-22, 02:30 #10 davar55   May 2004 New York City 108B16 Posts This is simpler as a double check: The #triangles sequence is 1,4,16,64,256,1024,4096,... Cut off at 1024. This gives 5 steps. The lengths are 3, 3/2 * 3, 3/2 * 5, 3/2 *17, 3/2 * 33, 3/2 * 65, ... i.e. 3, 4.5, 7.5, 25.5, 49.5, 97.5, ... Cut off at five steps gives: 49.5.    2006-05-22, 17:01   #11
mfgoode
Bronze Medalist

Jan 2004
Mumbai,India

22·33·19 Posts Quote:
 Originally Posted by davar55 This is simpler as a double check: The #triangles sequence is 1,4,16,64,256,1024,4096,... Cut off at 1024. This gives 5 steps. The lengths are 3, 3/2 * 3, 3/2 * 5, 3/2 *17, 3/2 * 33, 3/2 * 65, ... i.e. 3, 4.5, 7.5, 25.5, 49.5, 97.5, ... Cut off at five steps gives: 49.5.  If Im not mistaken you are looking at the problem in retrospect.

If you had to start just at what is given i.e. 3 how would you arrive at 3/2 and the multipliers 3 , 5 , 17 , 33 and 65? ....
They dont seem to be related in any way to me. Can you kindly clarify?
Mally    Thread Tools Show Printable Version Email this Page Similar Threads Thread Thread Starter Forum Replies Last Post MattcAnderson Homework Help 12 2016-11-08 12:10 MattcAnderson Puzzles 2 2015-11-07 17:54 davieddy Puzzles 2 2010-06-29 13:18 davar55 Puzzles 4 2009-07-02 20:16 Zeta-Flux Puzzles 12 2007-03-16 19:05

All times are UTC. The time now is 06:35.

Mon Dec 6 06:35:32 UTC 2021 up 136 days, 1:04, 0 users, load averages: 1.61, 1.56, 1.52