Mathoverflow 75792 #
Various questions about integer complexity, which is the minimum number of 1s needed to express a natural number using addition, multiplication, and parentheses.
Let ‖n‖ denote the integer complexity of n > 0.
- It is known that
‖3^n‖ = 3nforn > 0. - Is it true that
‖2^n‖ = 2nforn > 0? - The corresponding conjecture for
5is false, because5^6 = 15625 = 1 + 2^3 * 3^2 * (1 + 2^3 * 3^3)!
We have chosen to formalise this using an inductive type.
References:
- mathoverflow/75792 by user Harry Altman
- http://arxiv.org/abs/1203.6462 by Jānis Iraids, Kaspars Balodis, Juris Čerņenoks, Mārtiņš Opmanis, Rihards Opmanis, Kārlis Podnieks
- http://arxiv.org/abs/1207.4841 by Harry Altman, Joshua Zelinsky
- https://oeis.org/A005245 : Mahler-Popken complexity.
@[irreducible]
Equations
- One or more equations did not get rendered due to their size.
- Mathoverflow75792.Reachable.decide 0 x✝ = isFalse ⋯
- Mathoverflow75792.Reachable.decide 1 0 = isFalse Mathoverflow75792.Reachable.decide._proof_1
- Mathoverflow75792.Reachable.decide 1 n.succ = isTrue ⋯
The (Mahler-Popken) complexity of n:
the minimum number of 1s needed to express a given number using only addition and
multiplication. E.g. 2 = 1 + 1, so complexity 2 = 2.
Equations
- Mathoverflow75792.complexity n = if h : n = 0 then 0 else Nat.find ⋯
Instances For
theorem
Mathoverflow75792.Reachable.complexity_eq
{m n : ℕ}
(h : Reachable m n)
(min : ∀ n' < n, ¬Reachable m n')
:
5^6 = 15625 = 1 + 2^3 * 3^2 * (1 + 2^3 * 3^3)!
Is 5n the complexity of 5^n for 0 < n? Answer: No.
Is 3n the complexity of 3^n for 0 < n? Answer: Yes, by John Selfridge.
Reference: https://arxiv.org/abs/1207.4841
Is 2n the complexity of 2^n for 0 < n?