Go back
Inchworm

Inchworm

Posers and Puzzles

P
Bananarama

False berry

Joined
14 Feb 04
Moves
28719
Clock
13 May 05
Vote Up
Vote Down

I was thinking about this last night (how sad is that!). I think it's a case of proportional growth of the worm's speed, which is why the continuous model increases so much faster than the discrete model.

Basically, the worm travels at a constant rate under its own power, and gets a speed boost from the rope being stretched. At the beginning of its trip the worm hardly gets any boost from the stretch, because the fraction of the rope that it has already traveled only gets a fractional increase in length. As the worm travels further down the rope, it gets more of a boost, which makes it travel down the rope faster, which increase the boost, etc... which ends up making the worm's speed increase exponentially.

This is analagous to proportional growth in a population (where the growth rate is proportional to the population size), or continuously compounded interest (where the monetary increase is proportional to the size of the bank account).

In the discrete model, the worm only gets a boost at certain intervals, so the boost is only contributing to the factor which controls the change in speed with time (the worm's position) part of the time.

Interesting, though. I think the continuously stretched rope gives the minimum time to traverse the rope.

P
Bananarama

False berry

Joined
14 Feb 04
Moves
28719
Clock
13 May 05
Vote Up
Vote Down

Oh, I found a typo in my solution. The very first equation (dL/dt = k) describes the movement of the end of the rope, not the worm's position.

d

Joined
04 Aug 01
Moves
2408
Clock
13 May 05
Vote Up
Vote Down

Originally posted by PBE6
Oh, I found a typo in my solution. The very first equation (dL/dt = k) describes the movement of the end of the rope, not the worm's position.
Yeah I saw that, but it doesn't change anything that follows, so I didn't see the need to be picky. With the actual substance of the analysis, I don't see anything wrong...

P
Bananarama

False berry

Joined
14 Feb 04
Moves
28719
Clock
13 May 05
Vote Up
Vote Down

Originally posted by davegage
Yeah I saw that, but it doesn't change anything that follows, so I didn't see the need to be picky. With the actual substance of the analysis, I don't see anything wrong...
What do you think about my other post? The one before the typo correction?

d

Joined
04 Aug 01
Moves
2408
Clock
13 May 05
5 edits
Vote Up
Vote Down

Originally posted by PBE6
What do you think about my other post? The one before the typo correction?
Okay, I am beginning to think that there may not be a discrepancy between the continuous and discrete cases:

Consider the following (sort of) general case:

Suppose the rope is initially k cm long, and we stretch it at a rate of k cm/sec. The worm speed is w cm/sec.

Then, by your continuous model, we can find the time it takes the worm to traverse the rope by setting W/L = 1, and by noting that in this case, L = L0 + kt, and L0 = k.

Doing this, I get the traverse time (given as N number of seconds) to be

N = exp(k/w) - 1.

Now consider the discrete model: according to it, the fraction of rope traversed after the Nth second is given by

f(N) = (w/k)[S(N)]

where S(N) is just the Harmonic series taken out to the Nth term. A cool little fact is that for large N, the following approximation applies

S(N) ~ ln(N) + &,

where & is the "Euler Constant" and is approximately 0.5772156649...

Using this handy approximation, we can solve for the time it takes the worm to traverse by simply setting f = 1: Upon doing this, I get that the discrete model predicts

N = exp[(k/w) - &].

Comparing the two, you find there is virtually no difference between them for large k/w, and I imagine that the discrepancy between the two can be explained in the nature of the approximation we used for S(N).

Thus I don't think the discrete and continuous models necessarily conflict.

Does any of this make sense?

FYI, you can find more information regarding the approximation I used here:

http://www.jimloy.com/algebra/hseries.htm

or here:

http://mercury.easternct.edu/s/shamaj/web/projects/harmonicSeries.pdf

Cookies help us deliver our Services. By using our Services or clicking I agree, you agree to our use of cookies. Learn More.