Steps Of Increasing Size

Starting from the origin, I make a sequence of moves such that, at the k th k^\text{th} move, I move either + k +k units or k -k units along the real line. Let m n m_n be the least number of moves required to reach a positive integer n , n, then evaluate lim n m n n . \lim_{n\rightarrow \infty} \frac{m_n}{\sqrt{n}}.


The answer is 1.414.

This section requires Javascript.
You are seeing this because something didn't load right. We suggest you, (a) try refreshing the page, (b) enabling javascript if it is disabled on your browser and, finally, (c) loading the non-javascript version of this page . We're sorry about the hassle.

2 solutions

Mark Hennings
Mar 5, 2017

For any integer k N k \in \mathbb{N} with k 2 k \ge 2 , consider integers n n such that 1 2 k ( k 1 ) < n 1 2 k ( k + 1 ) \tfrac12k(k-1) < n \le \tfrac12k(k+1) . Since 1 + 2 + 3 + + ( k 1 ) = 1 2 k ( k 1 ) 1 + 2 + 3 + \cdots + (k-1) \; = \; \tfrac12k(k-1) we deduce that m n > k 1 m_n > k-1 , since n n is bigger than the largest positive integer that can be reached in k 1 k-1 steps. On the other hand 1 + 2 + 3 + + k = 1 2 k ( k + 1 ) 1 + 2 + + ( j 1 ) + ( j ) + ( j + 1 ) + + k = 1 2 k ( k + 1 ) 2 j \begin{aligned} 1 + 2 + 3 + \cdots + k & = \tfrac12k(k+1) \\ 1 + 2 + \cdots + (j-1) + (-j) + (j+1) + \cdots + k & = \tfrac12k(k+1) - 2j \end{aligned} for any 1 j k 1 \le j \le k , and so it is clear that m n k m_n \le k for any 1 2 k ( k 1 ) < n 1 2 k ( k + 1 ) \tfrac12k(k-1) < n \le \tfrac12k(k+1) such that n 1 2 k ( k + 1 ) ( m o d 2 ) n \equiv \tfrac12k(k+1) \pmod{2} . On the other hand 1 + 2 + + k + ( k + 1 ) ( k + 2 ) = 1 2 k ( k + 1 ) 1 1 + 2 + + ( j 1 ) + ( j ) + ( j + 1 ) + + k + ( k + 1 ) ( k + 2 ) = 1 2 k ( k + 1 ) ( 2 j + 1 ) \begin{aligned} 1 + 2 + \cdots + k + (k+1) - (k+2) & = \tfrac12k(k+1) - 1 \\ 1 + 2 + \cdots + (j-1) + (-j) + (j+1) + \cdots + k + (k+1) - (k+2) & = \; \tfrac12k(k+1) - (2j+1) \end{aligned} for any 1 j k 1 \le j \le k , and hence m n k + 2 m_n \le k+2 for all 1 2 k ( k 1 ) < n 1 2 k ( k + 1 ) \tfrac12k(k-1) < n \le \tfrac12k(k+1) with n ≢ 1 2 k ( k + 1 ) ( m o d 2 ) n \not\equiv \tfrac12k(k+1) \pmod{2} . In other words k 1 < m n k + 2 1 2 k ( k 1 ) < n 1 2 k ( k + 1 ) k-1 \; < \; m_n \; \le \; k+2 \hspace{2cm} \tfrac12k(k-1) < n \le \tfrac12k(k+1) and hence 2 ( k 1 ) k ( k + 1 ) < m n n < 2 ( k + 2 ) k ( k 1 ) 1 2 k ( k 1 ) < n 1 2 k ( k + 1 ) \frac{\sqrt{2}(k-1)}{\sqrt{k(k+1)}} \; < \; \frac{m_n}{\sqrt{n}} \; < \; \frac{\sqrt{2}(k+2)}{\sqrt{k(k-1)}} \hspace{2cm} \tfrac12k(k-1) < n \le \tfrac12k(k+1) Letting k k \to \infty we deduce that lim n m n n = 2 \lim_{n \to \infty} \frac{m_n}{\sqrt{n}} \,=\, \boxed{\sqrt{2}} .

@Mark Hennings Yes I did the same. I had been working on a similar problem with squares and no good method has struck me yet. Probably a CS problem.

Kartik Sharma - 4 years, 3 months ago

Sir, is there any method by which we can find a closed form for m(n) ?

Indraneel Mukhopadhyaya - 4 years, 3 months ago

Log in to reply

You can follow through the solution, make the minor adjustment below, and determine a closed form for m ( n ) m(n) . It would be of the form m n = 2 n + 1 2 + 0 , 1 , 2 m_n = \lfloor \sqrt{ 2n } + \frac{1}{2} \rfloor + 0, 1, 2 depending on certain circumstances.

More explicitly, the crux of the argument is

  1. The bound on the total sum (which gives us k)
  2. The parity of the sum (which tells us if we need an additional +1 or +2).

Minor adjustment: When k + 1 k+1 is odd, we do not need to use ( k + 1 ) ( k + 2 ) = 1 (k+1) - (k+2) = 1 to force an odd term.

Calvin Lin Staff - 4 years, 3 months ago
Kartik Sharma
Mar 9, 2017

The equivalent generating function of "in k k th move, I move either + k +k or k -k " is -

f n ( x ) = k = 1 n ( x k + x k ) f_n(x) = \prod_{k=1}^n {\left(x^k + x^{-k}\right)}

such that if we want to reach the number N N , then it must be the coefficient of this function, and we have to find minimum such n n .

f n ( x ) = 1 x n ( n + 1 ) / 2 k = 1 n ( 1 + x 2 k ) f_n(x) = \frac{1}{x^{n(n+1)/2}}\prod_{k=1}^n {\left(1 + x^{2k} \right)}

f n ( x ) = 1 x n ( n + 1 ) / 2 ( 1 + x 2 + x 4 + 2 x 6 + + a n x n ( n + 1 ) ) f_n(x) = \frac{1}{x^{n(n+1)/2}} \left(1 + x^2 + x^4 + 2x^6 + \cdots + a_n x^{n(n+1)}\right)

This clearly shows us that we have to minimum n n such that N n ( n + 1 ) 2 0 N - \frac{n(n+1)}{2} \geq 0

n 2 + n 2 N 0 n^2 + n - 2N \geq 0

n 1 + 1 + 8 N 2 n \geq \frac{-1 + \sqrt{1 + 8N}}{2}

Hence, lim n m n n = lim n 1 + 1 + 8 N 2 n = 2 \displaystyle \lim_{n\rightarrow \infty} \frac{m_n}{n} = \lim_{n\rightarrow \infty} \frac{\frac{-1 + \sqrt{1 + 8N}}{2}}{n} = \sqrt{2}

@Mark Hennings Can you give me an advice if we can follow this method with generating function for ± 1 2 ± 2 2 ± 3 2 ± 4 2 ± ± n 2 \pm 1^2 \pm 2^2 \pm 3^2 \pm 4^2 \pm \cdots \pm n^2 ?

I have followed it quite much and have made a "not-so good looking" recurrence on coefficients which can help make a computer program but not any good mathematical analysis that I know.

What I really would like to know is what will it look like at infinity.

Kartik Sharma - 4 years, 3 months ago

Log in to reply

I could also obviously think about the trigonometric analogy but haven't given it a lot of thought.

Kartik Sharma - 4 years, 3 months ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...