Page 1 of 1

Finite number of steps (own)

Posted: Wed Jan 05, 2011 8:35 pm
by nayel
Start with any real number other than $0$ and $1$. At each step, replace the number $x$ by either $1/x$ or $1-x$. Prove that, if we can get from $a$ to $b$ in a finite number of steps, the maximum number of steps required is two.

Example:
$3\xrightarrow{1/x} 1/3\xrightarrow{1-x} 2/3\xrightarrow{1/x} 3/2\xrightarrow{1-x} -1/2$, number of steps $=4$.
But $3\xrightarrow{1-x} -2\xrightarrow{1/x} -1/2$, number of steps $=2$.

Re: Finite number of steps (own)

Posted: Wed Jan 05, 2011 9:37 pm
by Avik Roy
This problem is an interesting one. Though I think the maximum number of steps needed is 3.
It is trivial to show that any of the steps, if repeated in succesive terms, leaves us with the original state. The following reversible sequence shows us the rest:

$ a \xrightarrow{1-x} 1-a \xrightarrow{1/x} 1/(1-a) \xrightarrow{1-x} a/(a-1) \xrightarrow{1/x} (a-1)/a \xrightarrow{1-x} 1/a \xrightarrow{1/x} a$

Re: Finite number of steps (own)

Posted: Wed Jan 05, 2011 10:51 pm
by anumoshsad
Nice problem and nice solution.
I also think the number of steps needed is 3.
Let $f(x)=\dfrac{1}{x}$ and $g(x)=1-x$.So $g\circ f= 1- \dfrac{1}{x},f\circ g=\dfrac{1}{1-x},f\circ g\circ f=g\circ f\circ g=\dfrac{x}{x-1}$.
And the fact that the set $\{id,f,g,g\circ f, f\circ g,f\circ g\circ f\}$ is closed under the composition of functions proves it all.

Re: Finite number of steps (own)

Posted: Wed Jan 05, 2011 11:50 pm
by nayel
Yes, I'm sorry, the number of steps required should be 3. Nice solutions! :)

Try this slightly modified version I just posted: viewtopic.php?f=23&t=285

Re: Finite number of steps (own)

Posted: Fri Jan 07, 2011 6:59 pm
by abir91
সুন্দর সমস্যা এবং সমাধান। কিন্তু maximum এর জায়গায় minimum হবে মনে হইতেছে। :|