Community

Want to learn how to tackle those tough MATHCOUNTS and AMC counting and probability problems? Check out Art of Problem Solving's Introduction to Counting & Probability by David Patrick.
Login Register Memberlist Search AoPS Blogs Contests Galleries Forum Index
The time now is Fri Dec 04, 2009 7:55 am
All times are UTC - 8
View posts since last visit
View unanswered posts
Liouville by Maximum Modulus Principle
Moderators: College Playground Moderators
Post new topic   Reply to topic View previous topicView next topic
6 Posts • Page 1 of 1
Author Message
cn2_71828182846
P versus NP
P versus NP

Offline
Joined: 23 Jun 2004
Posts: 34
Location: USA

To rate posts you must be logged in
#1
Liouville by Maximum Modulus Principle

Prove Liouville's theorem (i.e. a bounded entire function is constant) using the maximum modulus principle. That is, DON'T use the Cauchy estimate.

PostPosted: Sun Dec 12, 2004 3:23 pm  Back to top 
  ProfilePM
Kent Merryfield
Birch & Swinnerton Dyer
Birch & Swinnerton Dyer

Offline
Joined: 11 Jun 2004
Posts: 11419
Location: Long Beach, CA
United States

To rate posts you must be logged in
#2
Note that the maximum modulus principle alone is insufficient. You must use at least some other property of holomorphic functions.

Let M be the set of smooth functions from \mathbb{C} to \mathbb{C} that satisfy the maximum modulus property: that is, if f\in M, then |f| has no local maxima. The (entire) holomorphic functions form a subset of M as do also the harmonic functions. Both the holomorphic and harmonic functions do satisfy Liouville's Theorem. But f(x+iy)=\tan^{-1}x+i\tan^{-1}y is also in M and it's bounded and not constant.

As for the proof itself, let f be a bounded holomorphic function, and consider g(z)=f\left(\frac1z\right). This a bounded function with an isolated singularity at 0. We make use of the theorem that says such a singularity must be removable. Thus, \lim_{|z|\to\infty}f(z) exists, and we can uniquely extend f to a continuous function on the one-point compactification of \mathbb{C} (the "Reimann sphere.")

That makes this newly extended version of f a continuous function on a compact set, so its modulus must assume a maximum. But if f is not constant, this would be a local maximum of |f| (or |g|), which is not permitted by the maximum modulus principle.

We do need that theorem on removable singularities.

PostPosted: Mon Dec 13, 2004 9:17 am  Back to top 
  ProfilePM
cn2_71828182846
P versus NP
P versus NP

Offline
Joined: 23 Jun 2004
Posts: 34
Location: USA

To rate posts you must be logged in
#3
How about this proof:

Let f be entire, and define g(z) = (f(z) - f(0))/z, which is also entire. Since f is bounded, as |z| goes to infinity, |g(z)| goes to 0. Thus we can choose R > 0 s.t. there exists :varepsilon: s.t. |g(z)| < :varepsilon: for all |z| > R and the maximum value of |g(z)| for |z| <= R is > :varepsilon: . Such a maximum exists because |g(z)| is continuous and the closed disc is compact. Let w = z be the maximum of |g(z)| in this closed disc. Then |g(w)| >= |g(z)| for all z in C. Thus by the global maximum modulus principle, since C is open, g must be constant. Since |g(z)| goes to 0 as |z| goes to infinity, g must be identically 0. Thus f(z) = f(0) for all z in C, whereby f is constant.

PostPosted: Mon Dec 13, 2004 5:57 pm  Back to top 
  ProfilePM
Kent Merryfield
Birch & Swinnerton Dyer
Birch & Swinnerton Dyer

Offline
Joined: 11 Jun 2004
Posts: 11419
Location: Long Beach, CA
United States

To rate posts you must be logged in
#4
That works, too. I prefer cn2_71828182846's proof.

Note that that proof also makes use of a removable singularity, but the fact that it is removable is more elementary than the version I used.

PostPosted: Mon Dec 13, 2004 6:21 pm  Back to top 
  ProfilePM
Myth
Birch & Swinnerton Dyer
Birch & Swinnerton Dyer


Offline
Joined: 02 Sep 2003
Posts: 4490
Location: Chelyabinsk, Russia
Russian Federation

To rate posts you must be logged in
#5
Very nice proof, cn2_71828182846!
_________________
Myth is out of here

PostPosted: Mon Dec 13, 2004 10:16 pm  Back to top 
  ProfilePM
cn2_71828182846
P versus NP
P versus NP

Offline
Joined: 23 Jun 2004
Posts: 34
Location: USA

To rate posts you must be logged in
#6
Thanks!

PostPosted: Tue Dec 14, 2004 8:39 pm  Back to top 
  ProfilePM
Display posts from previous:   Sort by:   
6 Posts • Page 1 of 1
Post new topic   Reply to topic View previous topicView next topic
Jump to:  

You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
You cannot attach files in this forum
You can download files in this forum
You cannot post calendar events in this forum


© Copyright 2008 AoPS Incorporated. All Rights Reserved. • FoundationPrivacyContact Us