ScienceLunchTalk Beamer
ScienceLunchTalk Beamer
Zhu Chengbo
Department of Mathematics
National University of Singapore
February, 2006
Outline
Dn = {r 2k π , sr 2k π | 0 ≤ k ≤ n − 1}.
n n
Characterizing properties:
e (identity transformation)
Two symmetries can be composed (operated one after
another) to obtain another symmetry: x ◦ y
A symmetry has an inverse which is also a symmetry: x −1
Symmetry compositions are “associative":
(x ◦ y ) ◦ z = x ◦ (y ◦ z)
More examples:
(A) In Rn : the group of linear transformations preserving the
distance
x12 + · · · + xn2 ,
denoted by O(n). This is also the group of symmetries of the
(n − 1)-dimensional sphere in Rn :
S n−1 = {x ∈ Rn : x12 + · · · + xn2 = 1}
R2n = R 2
· · ⊕ R2}
| ⊕ ·{z
n
(classical groups)
ρg : X
More examples:
(i) SL(2, R) acts on the upper half plane by fractional linear
transformations:
az + b a b
g◦z = , g=
cz + d c d
(ii) SL(2, R) acts on two dimensional unit disk and on the circle
(the boundary of the disk).
G y G/H
(Homogeneous spaces for G)
(i) and (ii) are particular examples of homogeneous spaces for
SL(2, R).
Basic idea: You can tell a lot about the real nature of G by
knowing what possible places G can appear (as a group of
transformations), namely through its actions.
Out of all actions, linear actions (by invertible linear
transformations) are by far the simplest.
All actions can be converted in some sense to linear
actions by the following scheme:
If
G y X,
let
C(X ) = space of functions on X .
(think C(X ) as the full collection of observables on X )
Then G acts on C(X ) by:
(g · F )(x) = F (g −1 · x), g ∈ G.
Big advantages:
Representations can be superimposed (direct sum).
Representations can multiplied (tensor product).
Elements in G are represented by invertible linear
transformations and so can be analyzed in great detail.
The family of invertible linear transformations must behave
(according to the symmetry law of G), thus greatly limiting
their possibilities.
eiθ 7→ eik θ .
Remarks:
Linear transformations having simultaneous eigenspace
decomposition commute with each other.
One therefore should not expect such such simultaneous
eigenspace decompositions for groups which are not
“commutative".
irreducible representations
−(1+νi) ax + b a b
(g · F )(x) = |cx + d| F( ), g −1 = .
cx + d c d
Fundamental tasks:
Find all possible irreducible representations of a given G
(done by Langlands)
Find all possible irreducible unitary representations of a
given G (only done for some Lie groups of low rank, like
SL(2, R))
For all naturally occurring representations, carry out the
harmonic synthesis
Hidden Symmetries!
(by certain action of Sp(2n, R))
My talk ends
Thank you!