Puzzle
3 minute read

Ponder This Challenge - May 2021 - Fibonacci-like sequence with no primes

The Fibonacci sequence is defined by F0=0,F1=1F_0 = 0, F_1 = 1 and the recurrence relationship Fn=Fn1+Fn2F_n = F_{n-1} + F_{n-2}.

We call any sequence A0,A1,A2,A_0, A_1, A_2,\ldots Fibonacci-like if it satisfies the recurrences An=An1+An2A_n = A_{n-1}+A_{n-2} for all n2n \ge 2.

The Fibonacci sequence contains many prime numbers. For instance, F3=2,F11=89F_3 = 2, F_{11} = 89, etc. Our goal is to see how to generate a fibonnaci-like sequence without any prime numbers and relatively prime initial elements. This amounts to finding suitable relatively prime A0A_0 and A1A_1.

The main step in the generation is finding a set [(p_1, m_1, a_1), (p_2, m_2, a_2),..., (p_t, m_t, a_t)] of triplets of the form (p_k, m_k, a_k) such that. 1akmk1 \le a_k \le m_k. 2. For every natural number n, we have that for some k, n is equivalent to aka_k modulo mkm_k (i.e. mkm_k divides nakn-a_k). 3. pkp_k is a prime divisor of the Fibonacci number FmkF_{m_k} 4. All the pkp_k are distinct (the mkm_k and aka_k can be non-distinct)

Given this set, one can generate A0A_0 and A1A_1 that satisfy1. A0A_0 is equivalent to FmkakF_{m_k-a_k} modulo pkp_k 2. A1A_1 is equivalent to Fmkak+1F_{m_k-a_k+1} modulo pkp_k

and one can prove that this sequence does not contain any primes by using the following easy-to-prove identity, which holds for any Fibonacci-like sequence: Am+n=AmFn1+Am+1FnA_{m+n} = A_mF_{n-1} + A_{m+1}F_n.

Your goal is to find the set [(p_1, m_1, a_1),..., (p_t, m_t, a_t)] of triplets satisfying conditions 1-4 described above. (Hint: A set of 18 elements exists where the only primes dividing its mkm_k elements are 2,3 and 5).

A bonus "*" will be given for computing A0A_0 and A1A_1 from the found set, and explaining why every element of AnA_n is divisible by one of the pkp_k elements, but not equal to it.


We will post the names of those who submit a correct, original solution! If you don't want your name posted then please include such a statement in your submission!

We invite visitors to our website to submit an elegant solution. Send your submission to the ponder@il.ibm.com.

If you have any problems you think we might enjoy, please send them in. All replies should be sent to: ponder@il.ibm.com

Solution

  • As many have pointed out, this problem was originally solved by Ronald Graham who found a specific set of triplets ( "A Fibonacci-like sequence of composite numbers" , also given in the excellent book "Concrete mathematics"). The challenge here was not copying the set from the paper!

    Alternative sets can also be given, such as
    s = [[2, 3, 3], [3, 4, 4], [5, 5, 5], [7, 8, 1], [17, 9, 1], [11, 10, 1], [61, 15, 2], [47, 16, 2], [19, 18, 7], [41, 20, 3], [23, 24, 14], [31, 30, 29], [2207, 32, 10], [107, 36, 22], [2161, 40, 13], [109441, 45, 4], [1103, 48, 26], [103681, 72, 70], [181, 90, 67], [769, 96, 58]]

    Yielding
    A0 = 44186851375096125029100906489607183486440
    A1 = 23990976680675392254644252020033167422281

Solvers

  • *Alper Halbutogullari (2/5/2021 9:15 AM IDT)
  • *Albert Stadler (2/5/2021 9:00 PM IDT)
  • *Lazar Ilic (2/5/2021 9:38 PM IDT)
  • Hakan Summakoğlu (2/5/2021 9:53 PM IDT)
  • *Bertram Felgenhauer (2/5/2021 11:48 PM IDT)
  • Reda Kebbaj (3/5/2021 1:06 AM IDT)
  • *Chu-Wee Lim (3/5/2021 5:17 PM IDT)
  • Andreas Stiller (4/5/2021 12:40 AM IDT)
  • *Victor Chang (4/5/2021 9:04 AM IDT)
  • Daniel Chong Jyh Tar (4/5/2021 5:54 PM IDT)
  • *Uoti Urpala (4/5/2021 7:25 PM IDT)
  • *Clive Tong (5/5/2021 1:33 PM IDT)
  • *Dieter Beckerle (5/5/2021 8:26 PM IDT)
  • Andrew Mullins (5/5/2021 11:51 PM IDT)
  • Walter Sebastian Gisler (6/5/2021 1:14 PM IDT)
  • Wang Longbu (7/5/2021 6:51 AM IDT)
  • Guillaume Escamocher (7/5/2021 2:26 PM IDT)
  • Alex Fleischer (7/5/2021 10:07 PM IDT)
  • Nir Drucker (9/5/2021 8:29 AM IDT)
  • Karl D’Souza (9/5/2021 6:27 PM IDT)
  • *Amos Guler (9/5/2021 7:01 PM IDT)
  • *James Muir (13/5/2021 7:17 AM IDT)
  • *Marco Bellocchi (15/5/2021 2:45 AM IDT)
  • Eran Vered (17/5/2021 2:14 PM IDT)
  • *Latchezar Christov (17/5/2021 5:32 PM IDT)
  • David Greer (17/5/2021 7:08 PM IDT)
  • Daniel Bitin (19/5/2021 11:27 PM IDT)
  • Liubing Yu (20/5/2021 5:53 PM IDT)
  • *Tim Walters (21/5/2021 3:20 AM IDT)
  • JJ Rabeyrin (22/5/2021 4:21 PM IDT)
  • Simeon Krastnikov (23/5/2021 6:43 PM IDT)
  • Lorenz Reichel (24/5/2021 11:37 AM IDT)
  • Li Li (27/5/2021 4:31 AM IDT)
  • Fabio Michele Negroni (27/5/2021 8:28 AM IDT)
  • *Hansraj Nahata (28/5/2021 12:28 AM IDT)
  • *Florian Fischer (29/5/2021 12:30 AM IDT)
  • Nyles Heise (29/5/2021 7:17 AM IDT)
  • *Reda Kebbaj (31/5/2021 12:13 AM IDT)
  • *Harald Bögeholz (31/5/2021 5:37 AM IDT)
  • Tamir Ganor & Shouky Dan (31/5/2021 5:13 PM IDT)
  • Todd Will (31/5/2021 9:18 PM IDT)
  • Reiner Martin (1/6/2021 12:19 AM IDT)
  • Radu-Alexandru Todor (1/6/2021 1:54 AM IDT)
  • *Motty Porat (2/6/2021 3:34 AM IDT)
  • *Oscar Volpatti (2/6/2021 9:01 AM IDT)
  • *Phil Proudman (4/6/2021 3:46 AM IDT)

Related posts