1
$\begingroup$

Two I had an interview, where they have asked me a physics puzzle. The question was,

If two balls have been dropped at the same time from same height. suppose Ball A will start bouncing by 10% of reducing its height from its very previous bounce. And Ball Bwill start bouncing by 20% of reducing its height from its very previous bounce. Then how many times they will cross each other?

I have answered him as,

Definitely, Ball B bounce less than Ball A. So, in this context

if Ball-B starts its bounce from height = x.

Then if I will apply this into a program,

int cnt = 0;//Pseudo Code
if(x != 0)
{
x = x * 20 / 100;
 cnt++;
}

So, final value of cnt = number of times they will cross each other.

But after this I was thinking that, this many not be correct answer as in first drop, the Ball-A might have bounced of height x + (x- (height bounced by Ball-B)). So, it is not necessary that they might have crossed each other.

Please me understanding this puzzle.

$\endgroup$
  • $\begingroup$ Surely they cross infinitely many times? $\endgroup$ – user7530 Oct 27 '13 at 15:25
  • $\begingroup$ @user7530 That was my first thought too, but I don't think it's right (see my answer for explanation.) $\endgroup$ – Trevor Wilson Oct 27 '13 at 17:39
2
$\begingroup$

Your code only measures how many times ball $B$ can bounce before the precision of the computer is unable to distinguish its height from zero. This seems like an entirely different question from the one you are asked!

A second point is that the question says (in my interpretation) that each bounce of ball $B$ is 80% of the height of the previous one. So in my interpretation ball $A$ is bouncier than ball $B$. You seem to agree with this at one point, put then your code suggests that each bounce of ball $B$ is 20% the height of the previous one.

Notice that each ball bounces infinitely many times, with the bounce times forming a convergent geometric series. This is because the time that a bounce takes to complete is proportional to the speed at the start of the bounce, which is reduced by a constant factor each time.

Denote the time of the first bounce by $t=0$. Let $t_A$ denote the sum of the bounce times for ball $A$, or equivalently the time at which $A$ completes all of its (infinitely many) bounces and begins sitting still. Similarly, let $t_B$ denote the sum of the bounce times for ball $A$, or equivalently the time at which $B$ completes all of its (infinitely many) bounces and begins sitting still.

Note that $t_B < t_A$, so ball $B$ has finished bouncing while ball $A$ has bounced only finitely many times. All of the crossings must therefore happen in the time interval $[0,t_B]$. Let $t_0 = 0, t_1,t_2,\ldots,t_n \le t_B$ list in increasing order the times at which ball $A$ hits the ground during this interval.

For each bounce $[t_i,t_{i+1}]$ of ball $A$, it crosses paths with ball $B$ at most twice: At most once on the way up (because at any given height it is moving faster than ball $B$ is) and at most once on the way down (for the same reason.) Therefore an upper bound on the number of crossings is given by $2n$.

Because ball $A$'s height in the middle of one of its bounces $[t_i,t_{i+1}]$ is higher than ball $B$ can ever be during that time interval, the only way that there could be fewer than two crossings during the time interval $[t_i,t_{i+1}]$ is in the "degenerate case" that ball $B$ is on the ground at time $t_i$ or at time $t_{i+1}$. This happens at the beginning of $A$'s first bounce at $t = 0$. I'll leave it up to you to calculate $t_B$, $n$, and $t_1,\ldots,t_n$, and to see whether this "degenerate case" ever happens again.

$\endgroup$
  • $\begingroup$ Ball A bouncing against the ground after ball B has come to rest doesn't count as a "crossing"? $\endgroup$ – user7530 Oct 27 '13 at 17:46
  • $\begingroup$ @user7530 No, I don't think it does. $\endgroup$ – Trevor Wilson Oct 27 '13 at 17:56
  • $\begingroup$ I don't think $t_B<t_A$ because $1/10\infty=1/5\infty$ in almost all mathematics. $\endgroup$ – Simply Beautiful Art Feb 5 '16 at 0:42
  • $\begingroup$ @SimpleArt Are you suggesting that $t_A$ and $t_B$ are infinite? They are sums of convergent geometric series. $\endgroup$ – Trevor Wilson Feb 5 '16 at 1:33
  • $\begingroup$ @TrevorWilson Yes, which are both infinitely long, regardless of convergence. $\endgroup$ – Simply Beautiful Art Feb 5 '16 at 13:03

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.