Author Topic: A little problem (Read 3469 times)

AngelFish · « **on:** August 06, 2011, 01:36:42 am »

Here's a little ditty related to some stuff I've been working on. This problem has managed to crash Mathematica, confuse Matlab, destroy my attempts at a symbolic solution, and cause Wolfram Alpha to throw a temper tantrum. In short, it's a nasty little bugger and I was wondering if anyone else wanted to take a crack at it:

Lim_N→∞ N/(2^^{⌊Log₂N⌋})

+10 Math points to anyone who can either solve it or demonstrate that there's no elementary solution

phenomist · « **Reply #1 on:** August 06, 2011, 04:06:28 am »

No limit kind sir.

Consider N = 2^k, for arbitrarily large k. then the answer is 1.

Now consider N = 2^k-epsilon, for arbitrarily large k. then the limit is 2.

[note: k = integer]

undefined limit.

(note: have floors, abuse boundary conditions

)

AngelFish · « **Reply #2 on:** August 06, 2011, 05:15:20 am »

I suppose that should have been pretty obvious

Thanks.

* Qwerty.55 hands over the promised +10 points

Author Topic: A little problem (Read 3469 times)

AngelFish

A little problem

phenomist

Re: A little problem

AngelFish

Re: A little problem