I don't know what methods are taught, but I would think that the techniques go beyond what is taught in calculus 3 (less time spent on "why" means more time spent on "what"), and also knowing how to use that software (probably MATLAB, R, possibly SAS, SPSS, MathCAD, Maple, or Mathematica) is an important skill itself.
How about a physics problem for a change?
Consider the following scenario:
A small and heavy metal ball (negligible air resistance) is attached to the end of a rod (of negligible mass) which other end is attached to an axis (with negligible friction) so that the ball can swing freely. In other words, it's a pendulum.
The ball starts at zero velocity from the position depicted in the picture. There's a mechanism in the rod which releases the ball exactly at the point depicted in the picture.
The observer is located at the initial position of the ball. The observer has no visual contact with the system, but he can detect the sound of the ball hitting the ground, and measure exactly how much time passes between the initial position and the sound of the ball hitting the ground reaching the observer.
Assume that g = 9.8 m/s^2, and the speed of sound = 340 m/s. As depicted in the picture, the altitude of the ball (and thus also the observer) from the ground in the initial position is 50 m.
Exactly 5 seconds pass between the initial position and the sound reaching the observer. The question is: How long is the rod (measured from the center of the axis to the center of the ball)?
Er, I guess it's assumed that the radius of the ball is negligible, isn't it?
<klmz> it reminds me of that people used to keep quoting adelikat's IRC statements in the old good days
<adelikat> no doubt
<adelikat> klmz, they still do
Joined: 4/11/2006
Posts: 487
Location: North of Russia :[
Observer is 50m above ground then? because sound will take longer to get there than in case he was standing on the ground.
EDT: It's already answered in problem, stupid me
The pendulum should release the ball at the same time regardless of its length. Additionally, the ball launches with a velocity perpendicular to the pendulum (thus, 20° up from the horizontal). We can use the first fact to determine the amount of time the ball spends in flight, and the second to determine how far (horizontally) it flew during that time. That gives us the horizontal velocity, with which we can solve for the vertical velocity using basic trig; that in turn gives us the ball's kinetic energy, which was derived from its potential energy, which depends on the length of the rod.
I don't feel like plugging in the numbers, but this should be the right approach.
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
Joined: 1/16/2008
Posts: 358
Location: The Netherlands
It only releases at a fixed angle, the moment in time of release is dependent of the pendulum's length... we do not know the height of release (because we do not know the length of the pendulum) and thus not the amount of kinetic energy
I think you need to solve:
(amount of time for sound to arrive) + (amount of time in flight) = 5s
first term depends on the distance, thus on the horizontal velocity, thus on the speed
second term depends on the vertical velocity, thus also depends on the (same) speed
hence we need to rewrite such that speed becomes our only variable
edit: after that we know the kinetic energy, which equals the energy gained from losing height (while on the pendulum)... from which we can calculate the height-difference between point-of-release and initial point. Then some trig to get the pendulum's length
When the ball is released, its speed is sqrt(2*r*g*cos(pi/9)). That can be converted to horizontal and vertical velocity components easily enough.
In order to calculate the time taken to reach the release point, I'd need to integrate sqrt(sin(x)) or sqrt(cos(x)), and I'm not sure how to do this.
Joined: 11/30/2008
Posts: 650
Location: a little city in the middle of nowhere
I think this is how he got the formula:
E=mgh
m= 1 unit
g=9.8ms-2
h=r.cos pi/9
Kinetic energy, which is 100% of the energy is found by the formula
E=mv^2
m=1 unit
therefore
v=sqrt(2E)
E=g.r.cos(pi/9)
v=sqrt(2.g.r.cos(pi/9))
there, done
Joined: 4/20/2005
Posts: 2161
Location: Norrköping, Sweden
Hi, I have a question. I was thinking about this last night, but I couldn't come up with any answer:
Consider the function p(x) satisfying:
p(x) = 1 if x=p/q where p and q are both prime, p>=2, q>=2, and q>p.
p(x) = 0 else.
What I am wondering is, what properties does this function have on the interval D=[0,1]? Can the limit x->a p(x) be calculated at any point a in D? Is p continous in any point in D? What can be said about the set A={x: p(x)=1}? Is that a closed or open set? Because there are infinitely many primes, the set A will have infinitely many points, but I haven't been able to figure out much more than that.
Warp, I'm getting an answer of 16.285 meters. I'll explain how I did it later. Suffice to say, the fact that the large-angle pendulum is a nonlinear problem complicates things a lot, but it's still doable numerically.
A is countable and dense in [0,1]. This means that p(x) is nowhere continuous and its limit exists at no point. Additionally, it is not closed, as its closure is [0,1] =/= A, and it is not open, as clearly it cannot be expressed as the union of open intervals.
I now offer proof that A is dense in [0,1]. The fact that it is countable follows immediately from the fact that it is a subset of Q.
Let 1 > r1 > r2 > 0. We wish to find primes p and q such that r1*q > p > r2*q. The gap between our two limits, q*(r1-r2), grows linearly with q. However, for all sufficiently large primes, the gap between successive primes grows with O(p^(3/4)). Thus, by choosing q large enough, we can make q*(r1-r2) large enough that there must be a prime p between r1*q and r2*q. Thus, there exist p and q with r1 > p/q > r2 for all r1, r2 in [0,1], and thus A is dense in [0,1].
EDIT: The thought occurs to me that I should propose a problem. The infinitude of primes was brought up earlier. So let's look at another way to prove that fact.
a) The harmonic series is the sum of the reciprocals of all positive integers. Prove that this sum diverges. As a bonus point, show that it grows no slower than log(n).
b) Show that the divergence of the harmonic series necessarily implies that there is an infinite number of primes.
A few of you probably already know the answer to this one offhand. If you do, don't post it. This is one that's a lot more interesting to work out for oneself.
Joined: 4/20/2005
Posts: 2161
Location: Norrköping, Sweden
Very elegant solution! It seems p(x) has pretty much the exact same properties as Dirichlet's function. Thanks for the answer.
I have another question about primes. Many years ago I heard that there always exists at least one prime between n and 2*n, for all positive integers n>=2. Is this true, and if so, is there a proof for it?
Joined: 11/30/2008
Posts: 650
Location: a little city in the middle of nowhere
I have a shitty proof for this, but it works took 20 mins to solve.
The harmonic series is:
1/1 + 1/2 + 1/3 + 1/4 ...
if we remove all other terms form the series, and make it into a new series, we get
1/2 + 1/4 + 1/6 + 1/8 ...
then factorise out 1/2 we get
1/2 ( 1/1 + 1/2 + 1/3 + 1/4 ...
leaving behind:
1/1 + 1/3 + 1/5 + 1/7 ...
I'm going to subtract n from the odd terms. so it makes it nice:
1/2 + 1/4 + 1/6 + 1/8 ...
Basically I subtracted 1/2 from the first term, 1/12 from the second term, 1/30 from the next ect., making all of the denominators even.
wait a minute:
1/2 ( 1/1 + 1/2 + 1/3 + 1/4 ...
so if we add all the components together we get:
n + 1/2( 1/1 + 1/2 + 1/3 + 1/4 ...) + 1/2( 1/1 + 1/2 + 1/3 + 1/4 ...)
or in other words, the harmonic series is the sum of it's two halves, plus a finite number. Oh wow, That must mean It's divergent
The growth of f(x) where
f(x)= 1/1 + 1/2 + 1/3 + 1/4 ... 1/x
Is no less ln(x). By definition ln(x) is the area under the graph y=1/x starting at 1. f(x) is an approximation of this.
For every term in f(x), say there is a box with width 1 and height 1/x so the area of each box is equal to that term. Term 1 represents a box (1,1),(2,0) or generally (x , 1/x),(x+1 , 0) the top left corner (x , 1/x) passes through the line y=1/x. Although this isn't a formal proof, you can say that the area of the series of boxes is larger than the area underneath y=1/x because it covers the entire area underneath the graph plus the area outside where the top right corner is. Therefore f(x) > ln (x), x!=0
Because the harmonic series is made of converging series (all terms where the denominator is even for example. add them up, you will get a finite number), it must be made on an infinite number of converging series in order to diverge. for every converging series, there is a starting number (1/p) where p is prime, and yeh, you get the drift.
There is an easier one:
The third and fourth terms combined are greater than 1/4+1/4=1/2
The next four terms combined are greater than 1/8+1/8+1/8+1/8=1/2
The next 8 terms combined...
so the series
1/1+1/2+1/3+1/4+1/5+1/6+1/7+1/8+1/9+...
is greater than
1/1+1/2+1/4+1/4+1/8+1/8+1/8+1/8+1/16+...
which can be rearranged as
1+1/2+2*1/4+4*1/8+8*1/16+16*1/32+32*1/64+...
which simplifies to 1+1/2+1/2+1/2+1/2+1/2+1/2+...
which is divergent, and a positive series that is larger than a divergent positive series is itself divergent.
Your proofs both work, andymac, as does arflech's. arflech's leads to the non-calculus way to show the harmonic series grows no slower than log(n), as the comparison sequence clearly grows as log2(n).
Not quite correct. The sum of the reciprocals of the even numbers, or the sum of the reciprocals of any set of multiples of a number, is just as divergent as the harmonic series.
The reciprocals of prime numbers produce a divergent series. I wonder how hard it is to prove that... (I'm assuming it's probably very hard, as almost anything involving primes.)
Joined: 11/30/2008
Posts: 650
Location: a little city in the middle of nowhere
oh whoops, got it confused with a series like: 1/1 + 1/2 + 1/4 + 1/8 ect. I'm really just trying to confuse people with appalling grammar and incoherent sentences.
EDIT: still, there is the small issue of the sum of terms 1/(p^n) for any prime p, and whole n, which is convergent, for iterating n, but the sum of infinitely many of these could be divergent, corresponding to an infinite number of p, and if the remaining terms that are not of the form 1/(p^n) are a convergent series, then we have proven an infinite number of primes. However, I think the remaining terms would be divergent, and/or the sum of 1/(p^n) terms would be convergent, which would invalidate this theory.
andymac, you're on the right track here. But you are right: the reciprocals of composite numbers mess up your current argument.
I was actually getting to that...the proof that the harmonic series diverges implies an inifinite number of primes can be continued to not only show that the sum of the reciprocals of primes diverges, but also gives an estimate of its growth rate. This in turn, can be used to estimate the density of primes.
While the above method can be done rigorously, there are a few non-rigorous methods that still get the right answer and are still mostly correct.
Also, was my answer to your earlier problem (with the pendulum) correct?
I honestly apologize if I gave the impression that I have done the math myself and thus know the answer. I don't know the answer, nor how to calculate it (as you mentioned, pendulum math is very complicated and cannot be done analytically). I did choose values which I was pretty certain would give a rational answer (ie. the length of the rod would not be larger than the distance to the ground, etc), but I haven't solved the problem myself.
My sole intention was to try to come up with a relatively simple physical scenario which is nevertheless very complicated to solve. I was interested to see how the problem is solved, if someone knows.
Oh...well, in that case, I'll post how I solved it.
There are 3 parts: the swing, the flight, and the return of the sound. Each of these depends on L in a sufficiently simple way to make the problem numerically solvable.
1) when the pendulum has moved T from vertical, the height difference from its initial position is L*sin(T). Since energy should be conserved, m*g*dh=I*w^2/2 =mL^2*w^2/2=> w=sqrt(2*g/L^2*dh)=sqrt(2*g/L*sin(T))=w0*sqrt(2*sin(T)). So the angular velocity at angle T is w0*sqrt(2*sin(T)). Note that w0=sqrt(g/L) is the natural low-amplitude frequency of the pendulum.
now, as we know, w=dT/dt, which can be rearranged into dt=dT/w. integrating from T=0 to T=11pi/18 gives t=INT(dT/sqrt(sin(T)),0,11pi/18) / (sqrt(2)*w0). The integral is a constant, though it can only be analytically represented through elliptic functions. If we evaluate it numerically, we find...
t1=2.1034628 / w0
2) Now the ball has been released. For this part, S = sin(pi / 9) and C = cos(pi / 9). This will save space.
The ball has initial height 50 - LC, initial upward velocity L*w0*sqrt(2C)S, and of course, a downward acceleration of g. We want to know when it hits the ground. Putting these all together yields...
-.5g*t^2 + L*w0*sqrt(2C)S*t + 50-LC=0 => (w0*t)^2 - 2sqrt(2C)S*(w0*t)+ 2C + 100/L = 0
This is a simple quadratic equation, and can be solved quite easily. The answer, choosing the positive root, is
t2 = (sqrt(2C)S+sqrt(100/L-2C^3))/w0
3) The total horizontal displacement of the ball is L+LS+L*w0*sqrt(2C)*C*t2, while its total vertical displacement is 50. So, if sound travels at speed Cs, the time it takes to reach you is.
t3=sqrt((L+LS+L*w0*sqrt(2C)*C*t2)^2+50^2)/Cs
Final) The total time is, of course, t1+t2+t3. While this does not depend on L in a particularly simple way, it is simple enough for a numerical solution to t1+t2+t3=5. And the result you get is L=4.526m
(I must have made an error in my previous calculation)
Joined: 11/30/2008
Posts: 650
Location: a little city in the middle of nowhere
Made a new proof, this one actually works.
For the sake of our argument let's assume that there are a finite amount of primes. and that the number of primes is t. and that pv is the vth prime.
Composite numbers can be expressed as prime factors, and the sum of all the reciprocals of all whole positive integers will therefore be:
(correct me if I made any mistakes here)
The sum in the brackets always converges, and is therefore finite, and there are a finite number of products (t terms). Therefore this series is finite, and our initial assumption must be untrue: that there are a finite number of primes. Therefore there must be an infinite number of primes.
EDIT: also, maybe for that physics question there is more than one solution. I wouldn't doubt it.