wu :: forums (http://www.ocf.berkeley.edu/~wwu/cgi-bin/yabb/YaBB.cgi)
riddles >> putnam exam (pure math) >> Multiplicative function
(Message started by: NickH on Jan 30th, 2003, 6:36pm)

Title: Multiplicative function
Post by NickH on Jan 30th, 2003, 6:36pm
Here's one from the 1963 Putnam (though that's not where I originally found it) that I think is a bit easier than most Putnams...

Suppose {f(n)} is a strictly increasing sequence of positive integers such that:


    f(2) = 2
    f is multiplicative (m and n coprime => f(mn) = f(m)f(n))


Show that f(n) = n for every positive integer n.

Title: Re: Multiplicative function
Post by Icarus on Jan 30th, 2003, 7:42pm
Some Putnams aren't that hard. Particularly if you see the trick. This one is pretty straightforward though. Question (I don't know the answer yet): is it still true if you drop "strictly" from the wording?

Title: Re: Multiplicative function
Post by Pietro K.C. on Feb 4th, 2003, 3:30pm
A word on not-so-hard Putnams: is it me or is the 1985 one (http://www.unl.edu/amc/a-activities/a7-problems/putnam/1985.pdf) REALLY easy?

I'd like to see how other people did this one. My method, particularly the last steps, seem to me open to many alternatives (though I haven't found them :P). It is as follows:

Since {f(n)} is strictly increasing and f(2) = 2, it's clear that f(n) >= n for all natural numbers n. Also, if f(n0) = n0 for some n0, then f(n) = n for all n < n0 also.

Suppose that f(n) is not identically equal to n; then there must be a least n0 such that f(n0) > n0, with f(n) = n for all n < n0. As far as we know right now, this could be anything greater than 2.

Suppose this n0 is not a prime or a prime power; then there exist coprime a,b < n0 such that n0 = ab. However, this entails

f(n0) = f(ab) = f(a)f(b) = ab = n0.

We conclude that, if the set A = {n : f(n) > n} is not empty, its least element must be of the form pk, p a prime.

Suppose now that k is greater than 1. We then have

f(pk+p) = f(p(pk-1+1)) = f(p)f(pk-1+1) = p(pk-1+1) = pk+p.

This implies that f(pk) = pk, by the observation in the beginning. Hence, min(A) = p, a prime.

Now suppose that p > 3; then one of (p+1)/2, (p+3)/2 is odd, and we have

f(p+i) = f(2*(p+i)/2) = f(2)f((p+i)/2) = 2*(p+i)/2 = p+i

for the appropriate i. So the only choice left is min(A) = 3.

This is where we ad-lib, and where I would like to see alternatives. To show that f(3) = 3, I assumed that f(3) >= 4 and reasoned as follows:

f(21) < f(22) <=> f(3)f(7) < f(2)f(11) => 4f(7) < 2f(11) => 2f(7) < f(11) => f(14) < f(11); absurd!

So min(A) does not exist, and hence A is empty. We have f(n) = n for all natural numbers n.


Title: Re: Multiplicative function
Post by Icarus on Feb 4th, 2003, 6:20pm
Unfortunately I don't remember exactly what I did. It may be that I overlooked the problem with 3.

Concerning the 1985 Putnams. I thought it was easier than the 1984 test. I was ranked 58th in '85, whereas I was somewhere in the lower 100s in 1984 (the only other one I competed in.) Apparently not everyone agreed it was that much easier!

Title: Re: Multiplicative function
Post by Pietro K.C. on Feb 5th, 2003, 7:00am
I don't have the Putnams prior to 1985... everywhere I look online, it's 1985 onwards. So I mean the "easiness" in comparison with the later ones.

For 1985, I was able to do all but two (A6 and B6) of the questions in just over two hours! Not that I sat down with the purpose of timing myself, but the solutions just kept coming. After this burst, it took me another fifty minutes or so to do A6, but I admit B6 still has me stumped.

This is in stark contrast to what I have been capable of in the later Putnams - that's why I thought this one might have been consensually considered easier. :)

Title: Re: Multiplicative function
Post by BNC on Feb 5th, 2003, 8:30am
Pietro,

1980 - 1999 are here:
http://amath.colorado.edu/activities/Putnam/

1975-.. here:
http://www.math.niu.edu/~rusin/problems-math/

And the winner! : 1938-...
http://www.kalva.demon.co.uk/putnam.html

Title: Re: Multiplicative function
Post by NickH on Feb 9th, 2003, 6:38pm
Pietro, that's a characteristically ingenious solution.  My solution is at http://www.qbyte.org/puzzles/p042s.html

Icarus, I'm not sure whether the result follows if we remove "strictly" from the wording.  I can't derive a contradiction from assuming f(3) = 2, but I may be missing something.



Powered by YaBB 1 Gold - SP 1.4!
Forum software copyright 2000-2004 Yet another Bulletin Board