Millennium Prize Awarded For Perelman's Poincaré Proof

epee1221 writes "The Clay Mathematics Institute has announced its acceptance of Dr. Grigori Perelman's proof of the Poincaré conjecture and awarded the first Millennium Prize. Poincaré questioned whether there exists a method for determining whether a three-dimensional manifold is a spherical: is there a 3-manifold not homologous to the 3-sphere in which any loop can be gradually shrunk to a single point? The Poincaré conjecture is that there is no such 3-manifold, i.e. any boundless 3-manifold in which the condition holds is homeomorphic to the 3-sphere. A sketch of the proof using language intended for the lay reader is available at Wikipedia."

Whatever...

[Bentov]

It's not like wants the money or anything. He should at least take it and form a scholarship in his name. Jeez, the man is like a ./er, he lives with his mother.

Re:So will he accept?

[Obyron]

I can't take credit for finding this. Another Slashdotter was kind enough to link it the last time Perelman came up, but I found this to be very enlightening and illustrative of Perelman's personality as well as the whole Yau controversy. It's an article from the New Yorker co-written by Sylvia Nasar, who wrote the biography of John Nash, A Beautiful Mind. It contains what was at the time the only interview with Grigori Perelman, but I'm not sure if that's still true.

Annals of Mathematic: Manifold Destiny [newyorker.com]

Re:Well, sure

[JoshuaZ]

Ricci flow is an incredibly clever and sophisticated set of techniques. It is a very difficult technique to use and is by no means a "cheat code" for manifold questions. Most obviously, Ricci flow has been used with success to answer some aspects of the geometrization conjecture http://en.wikipedia.org/wiki/Geometrization_conjecture [wikipedia.org] but still leaves a lot. In order to have a truly good understanding of low-dimensional manifolds we are likely going to need some additional technique that has not yet been discovered.

Re:English Please

[Coryoth]

It's really all about classifying shapes. For two dimensional things this is pretty easy, at least as far as the topology goes: you need to know the curvature and "how many holes does it have" and that's it -- this is the whole topologist not knowing a coffee cup from a donut since they both have one hole and hence can be deformed one into the other (note that this is two dimensional because we are considering the 2-dimensional surface on the donut and coffee cup). In dimensions higher than two things start getting trickier because more bizarre configurations become possible. Perelman's work, which actually goes toward proving the rather more far reaching Geometrization Conjecture (due to Thurston), essentially lays out how you can classify all the different (from a topological point of view) shapes of things in three dimensions and higher.

What are the implications? Well, one reasonable question is: what is the topology of the universe like; what shape is the universe? Since the universe is a three dimensional manifold that turns out to be tricky. Perelman's work lays out the groundwork to be able to answer such a question.

from a mathematician

[l2718]

It's easier to explain the two-dimensional version, that is the version about surfaces. A mathematical surface is a kind of quilt: it's what you get from stitching together patches, each of which looks like a small piece of the plane. Just like with the quilt, if you bend or deform the surface it still is the same surface. Surfaces are completely "floppy".

Now, most real-life quilts are rectangular and have a boundary where they end, but you can also "close" the quilt by stitching the boundary back onto itself -- what you get is a "closed" surface. For example, you can stitch all the boundary together and get a sphere. Or you can stitch opposite sides together and get a "torus" -- the surface of a doughnut. You can also make more complicated quilts, which look like the joining of several doughnuts, i.e. a doughnut with several holes.

Next, one way that the sphere and doughnut-surface differ is that the latter has a hole. The way we express this is by looping a closed piece of string along the surface. With the sphere you can always slide the piece of string off the surface (we say that the sphere is "simply connected"), but with the torus you can run a loop of string along it in such a way that no deformation will allow you to take it off (we say the doughnut is "multiply connected").

Finally, the "2d Poincare conjecture" is the statement that the only simply connected closed 2d surface is the sphere. In other words, if you can't link a loop with your closed quilt then your quilt can be deformed to be a round sphere. A strong version of this was proved by Poincare, among others.

Now for the real "Poincare Conjecture" that was proved by Perelman, replace "2d" by "3d", so the quilt comes from stitching little cubes instead of little squares. The "closed and simply connected" assumptions are the same, and the conclusion is that the quilt is, up to deformation, the 3d sphere. It's much harder to visualize since now the quilt may not fit into regular 3d space. For example, the 3d sphere is what you get by stitching the whole boundary of the 3d cube together into one point (recall how we got a 2d sphere!) -- which is not something that fits into ordinary 3d space.