Seminar Sophus Lie

3 (1993) 149176
version of July 15, 1995
A Survey of Noncommutative Geometry
Methods for Group Algebras
Do Ngoc Diep
Abstract
In this survey we shall report about a K-theoretic approach to study
group algebras. Following the example of the group of ane transformations
of the straight line, the method consists of: 1. Construction of irreducible
group representations (orbit method, category O), 2. Decomposition of the
group algebra into a sequence of repeated extensions, and nally 3. Com-
putation of the extension invariants by the methods from noncommutative
geometry (KK-theory, cyclic theories).
1. Introduction
The main problem we are interested in is how to characterize groups and
their group algebras. For nite and compact groups this problem is well solvable
with complex representation theory. It seems to be quite dicult for locally
compact groups. We focus our attention, in this survey only on the topological
method of characterizing the group C*-algebras.
This material is the subject for the talks at Seminar C*-algebras at the
Universit at Heidelberg, SFB 288 of the Humboldt Universit at zu Berlin, Technische
Universit at zu Berlin and the Seminar Sophus Lie at the Technische Hochschule
Darmstadt.
Acknowledgements. It is a great pleasure for the author to express his deep
thanks to Prof. Dr. J. Cuntz for numereous helpful discussions and for hospital-
ity at the Mathematisches Institut der Universit at Heidelberg, where the author
learned the periodic cyclic homology. The author thanks sincerely Prof. Dr. Th.
Friedrich, Prof. Dr. K. H. Hofmann and Prof. Dr. R. Seiler for their
invitations to give talks at their seminars on the subject. The improvement of this
article has been realized during a stay of the author at the Technische Hochschule
Darmstadt. The author would like to express his sincere thanks to Prof. Dr.
K. H. Hofmann for the invitation and helpful discussions, and for the scientic
stimulation.
150 Diep
2. The Scope and an Example
2.1. The Problem
Let us rst of all consider a nite group G, [G[ < . It is easy to see that
the group G can be included in some (co-)algebras, more precisely some Hopf
bialgebra, for example in its complex group (Hopf bi-)algebra, G C[G],
g G

g
(g

)g

C[G],
which consists of the formal linear combinations of the form

gG
c
g
g, and where

g
is the Kronecker symbol corresponding to g. It is well-known that the group
representation theory of G is equivalent to the algebra representation theory of
C[G]. The last one is practically more exible to describe. With each representa-
tion of this group one considers the Fourier-Gelfand transform of the algebra
C[G] to the matrix algebra, corresponding to the representation ,

gG
c
g
g

gG
c
g
(g).
Let us denote

G the dual of G, i.e. the set of equivalence classes of irreducible
representations of G. We normally identify it with some set of representatives of
equivalence classes. It is well-known that:
a) The set

G is nite, i.e., there is only a nite number of nonequivalent
irreducible representations, say
1
, . . . ,
N
;
b) Each irreducible representation is nite dimensional, say of dimension n
i
, i =
1, 2, . . . , N;
c) This algebra C[G], by using the Fourier-Gelfand transform, is isomorphic
to the nite Cartesian product of matrix algebras
C[G]

=
N

i=1
Mat
n
i
(C).
This means that the structure of the group algebra C[G] and therefore of
the group G is well dened, if we obtain for G:
1) a good construction of all the irreducible representations
i
, i = 1, . . . , n of
G and
2) the Fourier-Gelfand transform, realizing the above cited isomorphism.
The problem is to extend this machinery to innite, say locally compact,
topological groups.
Let us from now on, consider a locally compact group G and some appro-
priate group algebras. The group algebra C[G] for G as an abstract group is not
enough to dene the structure of G. We must nd a more eective group algebra.
Diep 151
It is well known that for any locally compact group G one must in place of
the general linear representations consider the unitary ones. It is related to the fact
that in the general case one must consider also the innite dimensional represen-
tations, which are not always completely reducible. The unitary representations
however are completely reducible. In the case of a locally compact group G there
is a natural left-(right-)invariant Haar measure dg. The space L
2
(G) := L
2
(G, dg)
of the square-integrable functions plays an important role in harmonic analysis.
If the group is of type I, L
2
(G) admits a spectral decomposition with respect to
the left and right regular representations into a sum of the direct sum (the so
called discrete series) and/or the direct integral (the continuous series) of irre-
ducible unitary representations. The space L
1
(G) = L
1
(G, dg) of the functions
with integrable module plays a crucial role. With the well-dened convolution
product,
, L
1
(G) L
1
(G);
( )(x) :=
_
G
(y)(y
1
x)dy
it becomes a Banach algebra. There is also a well-dened Fourier-Gelfand trans-
formation on L
1
(G),
L
1
(G, g) ,
() := () =
_
G
(x)(x)dx.
If G is unimodular, one can also dene an involution

on L
1
(G) by

(g) := (g
1
).
There is a one-to-one correspondence between the (irreducible) unitary
representations of G and the nondegenerate (irreducible) *-representations of the
involutive Banach algebra L
1
(G). The general theorems of the spectral theory
of the representations of G are then proved with the help of an appropriate
translation into the corresponding theory for L
1
(G), for which one can use more
tools from functional analysis and topology. However the norm of the involutive
Banach algebra L
1
(G) is nonregular, i.e. in general
|a

a|
L
1
(G)
,= |a|
2
L
1
(G)
.
It is therefore more useful to consider the corresponding regular norm |.|
C

(G)
,
||
C

(G)
:= sup

G
|()|
and its completion C

(G). The spectral theory of unitary representations of G

is equivalent to the spectral theory of nondegenerate *-representations of the C*-
algebra C

(G). The general theorems of harmonic analysis say that the structure
of G can be completely dened by the structure of C

(G). One poses therefore the

problem of the description of the structure of the C*-algebras of locally compact
groups. This means that we must answer the questions:
1) How to realize the irreducible unitary representations of the locally compact
group G?
152 Diep
2) How to describe the image of the Fourier-Gelfand transform and in particu-
lar, of the inclusion of C

(G) into some continuous product of the algebras

L(H

),

G of bounded operators in the separable Hilbert space H

of
the representation .
To see that this is a good setting to extend the problem for nite group
to the class of locally compact groups, let us consider these questions for the
case of compact groups. Consider for the moment a compact group G. For a
compact group all irreducible representations are unitarizable, i.e., equivalent to
some unitary ones. It is also well-known that:
a) The set of equivalence classes of irreducible unitary representations is not
more than countable.
b) Each irreducible unitary representation is nite dimensional, say of dimen-
sion n
i
, i = 1, . . . , and there is some good realization of these representa-
tions, say in tensor spaces, or last time, in cohomologies.
c) The Fourier-Gelfand transformation gives us an isomorphism
C

(G)

i=1
Mat
n
i
(C).
This means that in the compact group case the group C*-algebra plays the
same role as the group algebra of a nite group.
Let us now return to the general case of a locally compact group.
The main problem is to describe the group algebra in general, and in par-
ticular the C*-algebra C

(G).
In general the problem of describing the structure of C*-algebras of non-
compact groups rests open up-to-date. This review outlines only the well-known
cases, where there is a nice interaction of the methods from noncommutative geom-
etry, say Orbit Method, category O, KK-theory, deformation quantization, cyclic
theories,. . . We restrict our attention mainly to Lie groups.
2.1.1. Analytic Methods
The rst nontrivial example is the group SL
2
(C). Its C*-algebra was studied by J.
M. G. Fell in 1961 in [F]. He described exactly the Fourier-Gelfand transforms
of C

(G) as some C*-algebra of sections of a continuous eld of operator algebras

over the dual. Many other mathematicians attemped to generalize his beautiful
but complicate analytic result to other groups. Nevertheless, until the moment
the only groups, the structure of whose C*-algebras were described explicitely are:
the abelian or compact groups and a few semi-simple Lie groups, say SL
2
(R) and
its univeral covering

SL
2
(R), the de Sitter group Spin(4, 1) and recently a family
G(p, q, ) of two step solvable Lie groups (see [De], [F], [M], [KM], [BM] and [W1]).
A fair amount is known about the C*-algebras of nilpotent Lie groups (see [P]),
including the Heisenberg groups. The C*-algebra of the Euclidean motion group
was studied by B. Evans [E]. Also P. Green [Gr] proposed another analytic
Diep 153
method for studying the C*-algebras of several solvable Lie groups. The results
are given very slowly and spectacularly.
One therefore has to develop another method, say to obtain some topological
invariants, which will be described in the rest of this paper.
2.1.2. The K-Theory Approach
The very useful K-functor for our approach is the operator KK-functor of G. G.
Kasparov [K1], generalizing the BDF K-functor [BDF], which characterizes the
isomorphy classes of short exact sequences of C*-algebras.
We are trying to decompose our C*-algebras into some towers of ideals and
step-by-step dene the associated extensions by KK-functors or their generaliza-
tions. The resulting invariants form just our index. This idea was proposed in
[D1] and developed in [D2] for a large class of type I C*-algebras. Hence, there
are two general problems:
(1) Find the C*-algebras which can be characterized by the well-known K-
functors, say by the operator K-functors.
(2) Generalize the theory of K-functors in such a way that they are applicable
for a larger class of C*-algebras.
Concerning the rst problem, we propose in [D8] a general construction
and some reduction procedure of the K-theory invariant Index C

(G) of group
C*-algebras. Using the orbit method [Ki], [D4] - [D7], we reduce Index C

(G)
to a family of A. Connes foliation C*-algebra indices Index C

(V
2n
i
, T
2n
i
),
see [C1]-[C2], by a family of KK-theory invariants. Using some generalization of
the Kasparov type condition (treated by G.G. Kasparov in the nilpotent Lie
group case [K2]), we reduce every Index C

(V
2n
i
, T
2n
i
) to a family of KK-theory
invariants of the same type valuated in KK(X,Y) type groups. The last ones are
in some sense computable by using the cup-cap product realizing the Fredholm
operator indices.
To demonstrate the idea, we consider the C*-algebra of the group of ane
transformations of the real straight line, but rst of all we need some new K-functor
tool. It is described in the next two subsections.
2.2. BDF K-Homology functor
Let us recall in this subsection the well-known BDF K-functor cxt. The main
reference is [BDF]. Denote by C(X) the C*-algebra of continuous complex-valued
functions over a xed metrizable compact space X, H a xed separable Hilbert
space over the complex numbers, L(H) and /(H) the C*-algebras of bounded and
respectively, compact linear operators in H. An extension of C*-algebras means
a short exact sequence of C*-algebras and *-homomorphisms of special type
0 /(H) c C(X) 0.
Two extensions are by denition equivalent i there exists an isomorphism :
c
1
c
2
and its restriction [
K(H
1
)
: /(H
1
) /(H
2
) such that the following
154 Diep
diagram is commutative
0 /(H
1
) c
1
C(X) 0

_[
.

_
_
_
_
0 /(H
2
) c
2
C(X) 0
There is a canonical universal extension of C*-algebras
0 /(H) L(H) /(H) 0,
the quotient algebra /(H)

= L(H)//(H) is well-known as the Calkin algebra.
By the construction of ber products, there is one-to-one correspondence between
the extensions of type
0 /(H) c C(X) 0
and the unital monomorphisms of type
: C(X) /(H).
Thus we can identify the extensions with the inclusions of C(X) into /(H).
Because all separable Hilbert spaces are isomorphic [Ki] and the automorphisms
of /(H) are inner and
Aut /(H)

= T|(H),
the projective unitary group, where |(H) denotes the unitary operator group,
we can identify the equivalences classes of extensions with the unitary conjugacy
classes of unital inclusions of C(X) into the Calkin algebra: Two extensions
1
and
2
are equivalent i there exists a unitary operator U : H
1
H
2
, such
that
2
=
U

1
, where by denition
U
: /(H
1
) /(H
2
) is the isomorphism
obtained from the inner isomorphism
U.().U
1
: L(H
1
) L(H
2
).
The extension : C(X) /(H) is called trivial i there exists a unital inclusion
: C(X) L(H) such that = , where : L(H) /(H) = L(H)//(H)
is the canonical quotient map. This inclusion corresponds to the split short
exact sequence. The sum of two extensions
i
: C(X) /
i
, i = 1, 2 is dened as
the extension

2
: C(X) /(H
1
) /(H
2
) /(H
1
H
2
).
This denition is also compatible with the equivalence classes of extensions. In
[BDF] the authors proved that:
1) The equivalence class of the trivial extension is the identity element with
respect to this sum.
2) For every metrizable compact space X, the set cxt
1
(X) of the equivalence
classes of extensions is an abelian group. One denes the higher groups by
cxt
1+n
(X) := cxt
1
(S
n
X), n = 0, 1, 2, . . . ,
Diep 155
3) cxt

is a generalized K-homology. In particular, the group cxt

1
(X) depends
only on the homotopy type of X and there is a homomorphism
Y

: c xt
1
(X) Hom
Z
(K
1
(X), Z)
which will be an isomorphism if X R
3
.
This K-homology is well developed and fruitfully applicable. It has many
applications in operator theory and in our problem of characterizing the group
C*-algebras. Let us demonstrate this in the rst example of the group of ane
transformations of the real straight line.
2.3. Topological Invariant Index
Let us in this subsection denote by G the group of all ane transformations of
the real straight line.
Theorem 2.1. Every irreducible unitary representation of the group G is uni-
tarily equivalent to one of the following mutually nonequivalent representations:
a) the representation S, realized in the space L
2
(R

,
dx
|x|
), where R

:= R (0),
and G acts according to the formula
(S
g
f)(x) = e

1bx
f(x), where g =
_
b
0 1
_
.
b) the representation U

, realized in C
1
and given by the formula
U

(g) = [[

1
(sgn )

, where R; = 0, 1.
Proof. See [GN].
This list of all the irreducible unitary representations gives the correspond-
ing list of all the irreducible nondegenerate unitary *-representations of the group
C*-algebra C

(G). In [D1] it was proved that

Theorem 2.2. The group C*-algebra with formally adjoint unit C

(G)

can
be included in a short exact sequence of C*-algebras and *-homomorphisms
0 / C

(G)

C(S
1
S
1
) 0,
i.e. the C*-algebra C

(G)

, following the BDF theory, is dened by an element,

called the index and denoted by Index C

(G)

, of the groups cxt

1
(S
1
S
1
)

= ZZ.
Proof. See [D1].
The innite dimensional representation S realizes the inclusion mentioned
above. Since
cxt(S
1
S
1
)

= Hom
Z
(
1
(S
1
S
1
), Z
1
)
is realized by a homomorphism from
1
(S
1
S
1
) to C

. Since the isomorphism

Y

: cxt(S
1
S
1
)

= Hom
Z
(
1
(S
1
S
1
), Z
1
)
156 Diep
is obtained by means of computing the indices and because the general type of
elements of
1
(S
1
S
1
) is g
k,l
= [g
0,1
]
k
[g
1,0
]
l
, k, l Z, we have
Ind (g
k,l
) = k Ind T(g
1,0
) +l Ind T(g
0,1
),
where T is the the *-isomorphism corresponding to S. Therefore it is enough
to compute the pair of indices Ind T(g
1,0
) and Ind T(g
0,1
). The last ones are
directely computed by the indices of the corresponding Fredholm operators.
Theorem 2.3.
Index C

(G) = (1, 1) cxt(S

1
S
1
)

= Z Z.
Proof. See [D1].
Let us now go to the general situation. To do this we must introduce
also some preparation about, rst of all, the construction of irreducible unitary
representations by means of the orbit method, then a method of decomposing the
C*-algebra into a tower of extensions and lastly to compute the index with the
help of the general KK-theory.
3. Multidimensional Orbit Methods
Let us in this section consider the problem of realization of irreducible uni-
tary representations of Lie groups. There are two versions of the orbit method; one
is the multidimensional quantization, the other is the innitesimal orbit method,
related with the so called category O.
3.1. Multidimensional Quantization
The orbit method can be constructed from the point of view of the theory of
holomorphically induced representations and also from the point of view of the
ideas of quantization from physics.
3.1.1. Construction of Partially Invariant Holomorphically Induced Representa-
tions
Let us consider a connected and simply connected Lie group G with Lie algebra
g := Lie(G). Denote by g
C
the complexication of g. The complex conjugation in
the Lie algebra will also be denoted by an overline sign. Consider the dual space
g

to the Lie algebra g. The group G acts on itself by the inner automorphisms
A(g) := g()g
1
: G G,
for each g G, conserving the identity element e as some xed point. It follows
therefore that the associated adjoint action A(g)

maps g = T
e
G into itself and
the coadjoint action K(g) := A(g
1
)

maps the dual space g

into itself. The orbit

space O(G) := g

/G is in general a bad topological space, namely non- Hausdor.

Consider one orbit O(G) and an element F g

in it. The stabilizer of

Diep 157
F is denoted by G
F
, its connected componenet by (G
F
)
0
and its Lie algebra by
g
F
:= Lie(G
F
). It is well-known that
G
F
G

F
is a principal bundle with the structural group G
F
. Let us x some connection in
this principal bundle, i.e. some trivialization of this bundle, see [SW]. We want to
construct representations in some cohomology spaces with coecients in the sheaf
of sections of some vector bundle associated with this principal bundle. It is well
known [SW] that every homogeneous vector bundle is an induced one with respect
to some representation of the structural group in the typical bre. It is natural to
x some unitary representation of G
F
such that its kernel contains (G
F
)
0
, the
character
F
of the connected component of the stabilizer is given by

F
(exp X) := exp (2

1 < F, X >)
and therefore the dierential D(
F
) = is some representation of the Lie
algebra g
F
. We suppose that the representation D(
F
) was extended to the
complexication (g
F
)
C
. The whole space of all sections seems to be too large for
the construction of irreducible unitary representations. One considers the invariant
subspaces with the help of some so called polarizations.
Denition 3.1. We say that a triple (p, ,
0
) is some ( , F)-polarization i :
a) p is some subalgebra of the complex Lie algebra g
C
, containing g
F
.
b) The subalgebra p is invariant under the action of all the operators of type
Ad
g
C
x, x G
F
.
c) The vector space p + p is the complexication of some real subalgebra m =
(p + p) g.
d) All the subgroups M
0
, H
0
, M, H are closed, where by denition M
0
(resp.,
H
0
) is the connected subgroup of G with the Lie algebra m (resp., h := pg)
and M := G
F
.M
0
, H := G
F
.H
0
.
e)
0
is an irreducible representation of H
0
in some Hilbert space V such that:
1. the restriction
0
[
G
F
H
0
is some multiple of the restriction
F
. [
G
F
H
0
,
where by denition
F
(exp X) := exp (2

1 < F, X >); 2. under the

action of G
F
on the dual

H
0
, the point
0
is xed.
f) is some representation of the complex Lie algebra p in V , which satises
the E. Nelson conditions for H
0
and [
h
= D
0
.
Let us recall that R. Blattner introduced the notion of mixed manifold
of type (k, l), see for example [Ki]. We consider the ber bundle, the base of
which is some type (k, l) mixed manifold and the bers of which are smooth m
dimensional manifold. We say that this ber bundle is some mixed manifold of
type (k, l, m).
158 Diep
Theorem 3.2. Suppose that
F
, , G
F
,etc. are as introduced above, and let
us denote by
F
the character of the group G
F
such that D
F
= 2

1F[
g
F
.
Then:
1) On the K-orbit
F
there exists a structure of some mixed manifold of type
(k, l, m), where
k = dimGdimM,
l =
1
2
(dimM dimH),
m = dimH dimG
F
.
2) There exists some irreducible unitary representation of the group H such
that its restriction [
G
F
is some multiple of the representation
F
. and
[
h
= D.
3) On the G-ber bundle c
|
G
F
= G
G
F
V associated with the representation
[
G
F
, there exists a structure of a partially invariant and partially holomor-
phic Hilbert vector G-bundle c
,
such that the natural representation of
G on the space of (partially invariant and partially holomorphic) sections
is equivalent to the representation by right translations of G in the space
C

(G; p, , F,
0
) of V -valued C

-functions on G satisfying the equations

f(hx) = (h)f(x), h H, x G,
L
X
f +(X)f = 0, X p,
where L
X
denotes the Lie derivative along the vector eld
X
on G, corre-
sponding to X.
Proof. The rst assertion is clear. The second one can be deduced from the
remark that the formula
(x, h) (I
V

F
. )(x).
0
(h)
denes an irreducible representation of the semidirect product G
F
H
0
which is
trivial on the kernel of the surjection
G
F
H
0
G
F
.H
0
.
This point is essential in the sense that with the assumption about the xed
point property of
0
we can ignore the Mackey obstruction, which appears when
we take the representations which are multiples of some representations
F
at
the restriction to some normal subgroup. M. Duflo [Du] considered two-fold
coverings to avoid this obstacle. See also [D5].
One can than apply the construction of unitarization c
,
in order to obtain
the corresponding unitary representation, which is noted by Ind(G; p, F, ,
0
).
One can dene also the representations in cohomologies with coecients in this
sheaf of partially invariant and partially holomorphic sections, which will be noted
by (L
2
Coh) Ind(G; p, F, ,
0
).
Diep 159
Remark 3.3. One introduces some order in the set of all ( , F)-polarizations
(p, ,
0
) (p

0
) p p

0
[
H
0

0
,

[
p
.
To have some irreducible resresentation, one must take the maximal polarizations
in this construction. It is interesting that this representations coincide with the
representations that appear via geometric quantization.
3.1.2. Multidimensional Geometric Quantization
Let us now consider the general concept of multidimensional geometric quantiza-
tion. Consider a symplectic manifold (M, ), i.e. a smooth manifold equipped
with a nondegenerate closed skew-symmetric dierential 2-form . The vector
space C

(M, ), with respect to the Poisson brackets

f
1
, f
2
C

f
1
, f
2
C

(M, )
becomes an innite dimensional Lie algebra.
Denition 3.4. A procedure of quantization is a correspondence associating to
each classical quantity f C

(M) a quantum quantity Q(f) L(H), i.e. a

closed, possibly unbounded, normal operator, which is autoadjoint if f is a real-
valued function, in some Hilbert space H, such that
Q(f
1
, f
2
) =
i

[Q(f
1
), Q(f
2
)],
Q(1) = Id
H
,
where := h/2 is the normalized Planck constant, and h is the unnormalized
Planck constant.
Let us denote by c a ber bundle of Hilbert spaces over M, a xed
connection preserving the Hilbert structure on the bers; in other words, if is
a curve connecting two points x and x

, the parallel transport along the way

provides a unitary isomorphism from the ber c
x
onto the ber c
x
. In this
case we can dene the corresponding covariant derivative

, Vect(M) :=
Der C

(M) in the space of smooth sections. One considers the invariant Hilbert
space L
2
(c
,
), which is the completion of the space (c
,
) of square-integrable
partially invariant and partially holomorphic sections.
Suppose from now on that M is a homogeneous G-space. Choose a triv-
ialization of the principal bundle G
x
G M, where G
x
is the stabilizer
of the point x on M. Let us denote by L

the Lie derivation corresponding to

the vector eld Vect(M). Let us denote by
1
(M) the connection form
of the ane connection on c . It is more comfortable to consider the normalized
connection form () =

1
(), the values of which are skew-adjoint operators
on bers. One has therefore

= L

(),
see for example [SW] for the nite dimensional case.
For each function f C

(M) one denotes

f
the corresponding Hamilto-
nian vector eld, i.e.
i(
f
) +df = 0.
160 Diep
Denition 3.5. We dene the geometrically quantized operator Q(f) as
Q(f) := f +

f
= f +

1
L

f
+(
f
).
Theorem 3.6. The following three conditions are equivalent:
1)
(, ) () () ([, ]) +

[(), ()] = (, ). Id
H
.
2) The curvature of the ane connection is equal to

(, ).Id
H
, i.e.
(, ) [

]
[,]
=

(, ).Id
H
.
3) The correspondence f Q(f) is a quantization procedure.
Proof. See [D6].
Suppose that the Lie group G acts on M by symplectomorphisms. Then
each element X of the Lie algebra g corresponds to a one-parameter subgroup
exp (tX) in G, which acts on M. Let us denote by
X
the corresponding strictly
hamiltonian vector eld. Let us denote also L
X
the Lie derivation along this vector
eld. We have
[L
X
, L
Y
] = L
[X,Y ]
,
and
L
X
f = f
X
, f.
Suppose that f
X
depends linearly on X. One has then a 2-cocycle of the
action
c(X, Y ) := f
X
, f
Y
f
[X,Y ]
.
Denition 3.7. We say that the action of G on M is at i this 2-cocycle is
trivial.
In this case we obtain from the quantization procedure a representation
of the Lie algebra g by the skew-adjoint operators
X

Q(f
X
)
and also a representation of g by the functions
X f
X
.
If the E. Nelson conditions are satised, we have a unitary representation of the
universal covering of the group G.
Diep 161
Theorem 3.8. The Lie derivative of the partially invariant and holomorphi-
cally induced representation Ind(G; p, F, ,
0
) of a connected Lie group G is just
the representation obtained from the procedure of multidimensional geometric quan-
tization, corresponding to a xed connection of the partially invariant partially
holomorphic induced unitarized bundle c
,
, i.e.
Lie
X
(Ind(G; p, F, ,
0
)) =

Q(f
X
).
Proof. See [D4].
Remark 3.9. The multidimensional version of the orbit method was developed
independently by the author in the language of multidimensional quantization [D4]
- [D7] and by M. Duo [Du], see also, [Ki] in the language of Mackeys method
of small subgroups. The result shows that for most of the connected Lie groups
the construction gives us at least a quantity of irreducible unitary representations,
enough to decompose the regular representations of G in L
2
(G), i.e. enough to
prove the Plancher`el formula [Du].
Remark 3.10. There are some reductions of this multidimensional quantiza-
tion procedure to the radical or nil-radical of stabilisors of type G
F
, see [DV], and
lifting them to U(1)-coverings [Vui1] - [Vui3], [Do1] - [Do2].
Remark 3.11. In [D10] the author proposed some method for common quan-
tization for foliations, the bers of which are the K-orbits, and its relation with
the integral Fourier operators.
3.2. Category O and globalization of Harish-Chandra modules
The construction of the irreducible unitary representations (L
2
Coh) Ind(G; p, F,
,
0
) in the Hilbert space L
2
(c
,
)

= L
2
(G; p, F, ,
0
) can in some cases be
considered as a globalization of a so called (g, K)-module, i.e. L
2
(c
,
0
)
(K)
itself
is some (g, K)-module, where K is some maximal compact subgroup of G. It is
therefore interesting to consider these (g, K)-module as some innitesimal version
of the orbit method. Let us see this in this subsection.
3.2.1. Admissible representations
Let us in this subsubsection recall some results about the Borell-Weil-Bott theorem
and the construction of admissible representations of nite dimensional semisimple
Lie groups as (g, K)-modules, [Wo], [M2].
If G is a compact connected Lie group, and F g

is a well-regular integral
functional on its Lie algebra, then the stabilizer is a maximal torus T , (If the
Harish-Chandra criterion for extistence of discrete series holds, it is a compact
Cartan subgroup.) and a choice of a positive system of roots
+
=
+
(g, t) denes
a G-invariant complex manifold structure on G/T in such a way that

+ g

represents the holomorphic tangent space. The character

, :=

F can
be extended to a character of the stabilizer G
F
= T , if the orbit is as usually
162 Diep
supposed to be integral. Let us denote in this case the induced bundle c
,
simply
by E

as in [Wo]. It is the associated homogeneous holomorphic hermitian line

bundle. One writes O(E

) G/T for the sheaf of germs of holomorphic sections

of E

G/T . The group G acts everywhere, including the cohomologies

H
q
(G/T; O(E

)). One denotes by :=

1
2

+ the half-sum of positive roots.

We cite from [Wo] the Borel-Weil-Bott theorem
Theorem 3.12. If + is singular then every H
q
(G/T; O(E

)) is trivial. If
+ is regular, let w denote the unique element of the Weyl group such that
w( +), ) > 0,
+
and let (w) denote its minimal length as a word in the simple root reections.
Then
i) H
q
(G/T; O(E

)) = 0 for all q ,= (w), and

ii) the action of G in H
(w)
(G/T; O(E

)) is the representation with heighest

weight w( +) .
This result was then extended for realizing the discrete series representa-
tions of general semisimple Lie groups. It is well known that one can induce from
these discrete series representations of a reductive part of parabolic subgroups to
obtain the tempered admissible representations of G. It was then remarked that
the representation of G in L
2
(c
,
) can be considered as the globalization of some
(g, K)-module , namely, L
2
(c
,
)
(K)
.
Lastly the tempered admissible representations are decribed in the coho-
mologies corresponding to T-modules [M2].
3.2.2. Discrete Series for loop groups
Let us now consider the loop groups associated with compact Lie groups. With
the help of the Zuckermanns derived functor, we can construct the innitesimal
version of the discrete series for loop groups, [D9]. The algebraic realization
of these representations are decribed in [D9] as a version of the Borel-Weil-Bott
and Kostant theorems. It is very interesting to develop a theory of tempered
representations for loop groups.
4. KK-theory
4.1. About KK-Functors
We now recall some essential points of Kasparovs setting of the KK-theory. It is
an analog of the Brown-Douglas-Fillmore theory, but settled for the general case.
4.1.1. Denitions
The main reference for this subsubsection is [JT]. Let A, B, E be C*-algebras,
and / the ideal of compact operators in some xed separable Hilbert space. Let
Diep 163
us consider the extensions of type
0 B / E A 0.
Two extensions are said to be equivalent i there is some isomorphism : E E

such that it induces the identity isomorphisms on the ideal B / and on the
quotient A, i.e. the following diagram is commutative
0 B / E A 0
_
_
_

_
_
_
_
0 B / E

A 0
The extension is called trivial if the the exact sequence splits. Also due to well-
known results of R. C. Busby, we can identify each extension with some *-
homomorphism from A to the algebra of exterior multipliers of B /, : A
O(B /). The sum of two extensions

i
: A O(B /)
can be therefore dened as the extension

2
: A O(B /) O(B /) O(B /) M
2

= O(B /),
where M
2
is the full algebra of 2 2-matrices over the complexes numbers. Two
extensions
i
, i = 1, 2 are stably equivalent if there exist two trivial extensions
1
and
2
such that the sums
i
+
i
, i = 1, 2 are equivalent.
G. G. Kasparov [K1] proved that:
i) when A is nuclear separable and B has at least an approximate identity, the
set Ext(A, B) of the stably equivalent classses of extensions is an ablelian
group.
ii) K

(.) = Ext

(A, .) is a K-homology theory and K

= Ext

(., B) is the
algebraic K-theory of C*-algebras.
iii) There is a natural realization of KK
,
as some K-bifunctor and for its direct
relation with the Ext

-groups see [K2].

J. Rosenberg and C. Schochet [RS] proved the K unneth formula for
these groups, i.e. there is some homomorphism
Y : Ext
i
(A, B)

j(mod 2)
Hom
Z
(K
i+j
(A), K
i+j+1
(B)).
Let us see this in more detail in the next subsubsection.
4.1.2. Relation with K-groups of C*-algebras
The most important for us is the relation of the theory with K-groups of C*-
algebras. Let A be an algebra with unit. By denition, K
0
(A) is the Grothendieck
group of the semi-group of the stably equivalent classes of projective A-modules
164 Diep
of nite type. When A has no unity element, one considers the algebra A

with
the formally adjoint unit and denes the K-group as
K

(A) := kerK

(A

) K

(C) = Z.
This denition is compatible with the above dened K-groups also for algebras
with unit element. For A = C(X), there is a natural isomorphism between these
K-groups with the corresponding topological groups K

(X), see for example [K].

One denes the higher groups K
n
(A) as
K
n
(A) := K
0
(AC
0
(R
n
)), n 0.
The Bott theorem says that K
0
(A)

= K
2
(A). The Connes-Kasparov theorem says
that for any connected and simply-connected solvable Lie group G,
K
0
(C

(G)) =
_
Z if dim G is even,
0 otherwise,
K
1
(C

(G)) =
_
0 if dim G is even,
Z otherwise.
For each extension
0 J E A 0,
there is a six-term exact sequence of K-groups
K
0
(E) K
0
(A)

K
0
(J) K
1
(J)

K
1
(A) K
0
(E)
Let us consider the case J = B/, There is an isomorphism between K

(J)
and K

(B). The group K

(A) consists of the formal dierences of equivalence

classes of projectors in A/. One obtain therefore the well-known exact sequence
K
0
(E) K
0
(A)

0
K
0
(B) K
1
(B)

1

K
1
(A) K
1
(E)
It is therefore clear that each element of Ext(A, B) induces a pair of homomor-
phisms (
0
,
1
) of K-groups, and one has a homomorphism
: Ext
i
(A, B)
jZ/(2)
Hom
Z
(K
i+j
(A), K
i+j+1
(B)),
associating to each extension a pair of connecting homomorphisms (
0
,
1
).
J. Rosenberg and S. Schochet [RS] have proved the following exact sequence
0

iZ/(2)
Ext
1
Z
(K
i+j
(A), K
i+j+1
(B)) Ext
i
(A, B)

jZ/(2)
Hom
Z
(K
i+j
(A), K
i+j+1
(B)) 0.
Diep 165
4.2. Construction and reduction of the K-Theory Invariant Index C

(G)
We review in this section a construction for obtaining the short exact sequence of
C*-algebras.
4.2.1. Measurable foliations
In this section we propose a canonical method for constructing the measurable
foliations, consisting of the adjoint orbits of xed dimension, and therefore their
C*-algebras [C1]. The last ones are included in group C*-algebras or their quo-
tients.
Let us denote by G a connected and simply connected Lie group, g = Lie(G)
its Lie algebra, g

= Hom
R
(g, R) the dual vector space, O = O(G) the space of
all the coadjoint orbits of G in g

. This space is a disjoint union of subspaces of

coadjoint orbits of xed dimension, i.e.
O = H
02ndimG
O
2n
,
O
2n
:= O; dim = 2n.
We dene
V
2n
:=
_
dim=2n
.
Then it is easy to see that V
2n
is the set of points of a xed rank of the Poisson
structure bilinear function
X, Y (F) = F, [X, Y ]).
suppose it is a foliation, at least for V
2n
, with 2n = max.
First, we shall show that the foliation V
2n
can be obtained by the associated
action of R
2n
on V
2n
via 2n times repeated action of R.
Indeed, xing any basis X
1
, X
2
, . . . , X
2n
of the tangent space g/g
F
of at
the point F , we can dene an action R
2n
V
2n
as
(R (R (. . . R V
2n
)))
by
(t
1
, t
2
, . . . , t
2n
) exp(t
1
X
1
) . . . exp(t
2n
X
2n
)F
Thus we have the Hamiltonian vector elds

k
:=
d
dt
k
[
t
k
=0
exp(t
k
X
k
)F, k = 1, 2, . . . , 2n
and the linear span
F
2n
=
1
,
2
, . . . ,
2n

provides a tangent distribution.

Theorem 4.1. (V
2n
, F
2n
) is a measurable foliation.
Proof. See [D8].
Corollary 4.2. The Connes C*-algebra C

(V
2n
, F
2n
), 0 2n dimG are well
dened.
166 Diep
4.2.2. Reduction of Index C

(G) to Index C

(V
2n
, F
2n
)
Now we assume that the orbit method (see [Ki],[D4]-[D6]) gives us a complete list
of irreducible representations of G ,

F
,
= Ind(G,
F
, , p), A
G
(F),
the nite set of Duos data.
Suppose that
O =
k
_
i=1
O
2n
i
is the decomposition of the orbit space as a stratication of orbits of dimensions
2n
i
, where n
1
> n
2
> . . . > n
k
> 0.
We include C

(V
2n
1
, F
2n
1
) into C

(G). It is well known that the Connes C*-

algebra of foliation can be included in the algebra of pseudodierential operators
of degree 0 as an ideal. This algebra of pseudodierential operators of degree 0 is
included in C

(G).
We dene
J
1
=

F
O(G)\O
2n
1
ker

F
,
,
and
A
1
= C

(G)/J
1
.
Then
C

(G)/C

(V
2n
1
, F
2n
1
)

= A
1
and we have
0 J
1
C

(G) A
1
0
Id
0 C

(V
2n
1
, F
2n
1
) C

(G) C

(G)/C

(V
2n
1
, F
2n
1
) 0
Hence J
1
C

(V
2n
1
, F
2n
1
) and we have
O C

(V
2n
1
, F
2n
1
) C

(G) A
1
0
Repeating the procedure in replacing
C

(G), C

(V
2n
1
, F
2n
1
), A
1
, J
1
by
A
1
, C

(V
2n
2
, F
2n
2
), A
2
, J
2
,
we have
0 C

(V
2n
2
, F
2n
2
) A
1
A
2
0
etc...
So we obtain the following result.
Diep 167
Theorem 4.3. The group C*-algebra C*(G) can be included in a nite sequence
of extensions
(
1
) : 0 C

(V
2n
1
, F
2n
1
) C

(G) A
1
0
(
2
) : 0 C

(V
2n
2
, F
2n
2
) A
1
A
2
0,
. . . . . . . . . . . . . . .
(
k
) : 0 C

(V
2n
k
, F
2n
k
) A
k1
A
k
0,
where

A
k
Char(G)
Corollary 4.4. Index C

(G) is reduced to the system Index C

(V
2n
i
, F
2n
i
), i =
1, 2, . . . , k by the invariants
[
i
] KK(A
i
, C

(V
2n
i
, F
2n
i
)), i = 1, 2, . . . , k.
Remark 4.5. Ideally, all these invariants [
i
] could be computed step-by-step
from [
k
] to [
1
].
4.2.3. Reduction of Index C

(V
2n
i
, F
2n
i
) to the computable extension indices with
values in topological KK-groups of pairs of spaces
Let us consider C

(V
2n
i
, F
2n
i
) for a xed i. We introduce the following assumptions
which were considered by G. G. Kasparov in nilpotent cases [K2]:
(A
1
) There exists k Z, 0 < k 2n
i
such that the foliation
V
gen
:= V
2n
i
(Lie )

has its C*- algebra

C

(V
gen
, F[
Vgen
)

= C(O

gen
) /(H),
where
:= R
k
R
2n
i
G,
Lie = R
k
g/g
F
i
, (Lie )

V
2n
i
.
Example 4.6. If V
gen
is a principal bundle, or the space O
gen
= V
gen
/G is a
Hausdor space, then C

(V
gen
, F[
Vgen
) C(O

gen
) /(H)
It is easy to see that if the condition (A
1
) holds, C

(V
2n
i
, F
2n
i
) is an
extension of C

(V
2n
i
V
gen
, F
2n
i
[
.
) by C(O

gen
)/(H), where O

gen
=

F
,
;
F

O
gen
, A
G
(F), described by the multidimensional orbit method from the
previous section. If k = 2n
i
, (R
2n
i
)

= 0, V
2n
i
= V
gen
, we have
C

(V
2n
i
, F
2n
i
) C(O

2n
i
) /(H)
If k = k
1
< 2n
i
, then R
2n
i
k
1
acts on V
2n
i
V
gen
and we suppose that a similar
assumption (A
2
) holds
168 Diep
(A
2
) There exists k
2
, 0 < k
2
2n
i
k
1
such that
(V
2n
i
V
gen
)
gen
:= (V
2n
i
V
gen
) (R
k
2
)

has its C*-algebra

C

((V
2n
i
V
gen
)
gen
, F
2n
i
[.) C((O
2n
i
O
gen
)
gen
)

/(H)
As above, if k
2
= 2n
i
k
1
, C

(V
2n
i
V
gen
, F
2n
i
[.) C((O
2n
i
O
gen
)

gen
) /(H).
In the other case we repeat the procedure and go to assumption (A
3
), etc. . . .
The procedure must be nished after a nite number of steps, say in the
m-th step,
C

((. . . (V
2n
i
V
gen
)(V
2n
i
V
gen
)
gen
. . . , F
2n
i
[
.
) C((. . . (O
2n
i
O
gen
). . .))/(H)
Thus we have the following result.
Theorem 4.7. If all the arizing assumptions (A
1
), (A
2
), . . . hold, the C*-algebra
C

(V
2n
i
, F
2n
i
) can be included in a nite sequence of extensions
0 C(O

gen
) /(H) C

(V
2n
i
, F
2n
i
) C

(V
2n
i
V
gen
, F
2n
i
[
.
) 0
0 C((O
2n
i
O
gen
)

gen
) /(H) C

(V
2n
i
V
gen
, F
2n
i
[
.
) C

(. . .) 0
. . . . . . . . . . . . . . .
0 C((. . . (O
2n
i
O
gen
) (O
2n
i
O
gen
))
gen
. . .

) /(H) C

(. . .)
C

(. . .) /(H) 0.
4.2.4. General remarks concerning computation of Index C*(G) via KK(X
i
, Y
i
)
We see that the general computation procedure of Index C*(G) is reduced to the
case of short exact sequences of type
() 0 C(Y ) /(H) c C(X) /(H) 0,
and the index is
[] = Index c KK(X, Y )
The group KK
i
(X, Y ) can be mapped onto

jZ/(2)
Hom
Z
(K
i+j
(X), K
i+j+1
(Y ))
with kernel

jZ/(2)
Ext
1
Z
(K
i+j
(X), K
i+j+1
(Y ))
by the well known cap-product, see [K2]. So [] = (
0
,
1
)

0
Hom
Z
(K
0
(X), K
1
(Y )) = Ext
0
(X) K
1
(Y )

1
Hom
Z
(K
1
(X), K
0
(Y )) = Ext
1
(X) K
0
(Y ).
Diep 169
Suppose e
1
, e
2
, . . . , e
n

1
(X) to be generators and
1
,
2
, . . . ,
n
c the corre-
sponding Fredholm operators, T
1
, T
2
, . . . , T
n
the Fredholm operators, representing
the generators of K
1
(Y ) = Index[Y, Fred] We have therefore
[
0
] =

j
c
ij
indexT
j
,
where

0
= (c
ij
) Mat
rank K
0
(X)rank K
1
(Y )
(Z)
In the same way
1
can be computed.
4.3. Case-by-Case Examples
We nish this section by an overview of well-treated examples, demonstrating
eectiveness of the above reduction and conputation.
Now we introduce the classes MD and MD4 of Lie groups, the Lie algebras
of which are well-dened by the method exposed above.
Denition 4.8. Lie algebra g is said to be in the class MD (resp., MD), i
for every F g

, dim
F
is either 0 or equal to dimg (resp., or maximal).
We nish this section by stating some current new research results, see
[Vu2],[R1],[Vu1],[SV].
Theorem 4.9. (Classication for MD and MD4). (1) The Lie algebra
g is of class MD if and only if either g is commutative or it is the Lie algebra
of the Lie group of ane transformations of the real straight line A R , and its
universal covering

A R , generated by two generators X, Y with the only non
trivial commutation relation
[X, Y ] = Y,
or the Lie algebra of the group of ane transformations of the complex straight
line A C and its universal covering

A C, generated by the generators T, X, Y, Z
with the only non trivial commutation relations
[T, X] = X, [T, Y ] = Y, [X, Y ] = Z, [X, Z] = Y
(2) Let us denote MD4 the subclass of MD , consisting only of the 4-
dimensional ones. Suppose that g is indecomposable into a direct product of two
proper ideals. Then g is of class MD4 if and only if it is generated by the
generators T, X, Y, Z with the only non-trivial commutation relations in one of
the following cases :
g
4,1,1
: [T, X] = Z,
g
4,1,2
: [T, Z] = Z,
g
4,2,1()
: [T, X] = X, [T, Y ] = Y ; R

= R (0),
170 Diep
g
4,2,2
: [T, X] = X, [T, Y ] = X +Y,
g
4,2,3()
: adT =
_
_
_
cos sin 0
sin cos 0
0 0 0
_
_
_,
g
4,2,4
= Lie(A C) : adT =
_
_
_
1 0 0
0 1 0
0 0 0
_
_
_, ad X =
_
_
_
0 1 0
1 0 0
0 0 0
_
_
_,
g
4,3,1(
1
,
2
)
: adT =
_
_
_

1
0 0
0
2
0
0 0 1
_
_
_,
1
,
2
R

,
g
4,3,2()
: adT =
_
_
_
1 0
0 0
0 0 1
_
_
_, R

,
g
4,3,3
: adT =
_
_
_
1 1 0
0 1 1
0 0 1
_
_
_,
g
4,3,4()
: adT =
_
_
_
cos sin 0
sin cos 0
0 0
_
_
_, R

, (0, ),
g
4,4,1
= Lie(R
J
H
3
) : adT =
_
_
_
0 1 0
1 0 0
0 0 0
_
_
_, [X, Y ] = Z,
g
4,4,2
= Lie(R H
3
) : ad T =
_
_
_
1 0 0
0 1 0
0 0 0
_
_
_, [X, Y ] = Z,
(in this case the group is the so called real diamond group).
Proof. See [Vu2].
It is not hard to classify all the K-orbits of these groups and then construct
the corresponding representations by the orbit method. Now we consider the C*-
algebras of the corresponding measurable foliations of generic orbits. Most of
them can be described by the analytic method via the corresponding Fourier-
Gelfand transforms, but we consider only the cases g
4,3,4(,)
= Lie(G
(,)
) ,
g
4,4,1
= Lie(R
J
H
3
) and the real diamond group R H
3
.
The structure of the C

-algebras of these examples is described in the

following results which we state, again without detailed proof, see [D1], [D2], [R1],
[R2] [Vu1], [Vu2].
Diep 171
Theorem 4.10. (1) The C*-algebra C

(A R) of the group of ane transfor-

mations of the real straight line R can be included in an C*-algebra extension
0 /(H) C

(A R) C(S
1
S
1
) 0
and its structure is uniquely dened by the K-theory invariant
Index C

(A R) = (1, 1) cxt(S
1
S
1
)

= Z Z.
(2) The C*-algebra C

((A R)
0
) of the group of proper ane transforma-
tions of the real straight line can be included in an extension
0 /(H) /(H) C

((A R)
0
) C(S
1
) 0
and its structure is uniquely dened by the K-theory invariant
Index C

((A R)
0
) = (1, 1) cxt(S
1
) cxt(S
1
) = Z Z
(3) The C*-algebra C

(A C) of the group of ane transformations of the

complex straight line C can be included in an extension
0 /(H) C

(A C) C(X) 0,
where
X := z C; [z 2
n
[ = 2
n
, n = 1, 2, ...
is the so called Hawaian necklace, and its structure is uniquely dened by the
K-theory invariant
Index C

(A C) = (1, 1, , . . .) KK(X, pt) = cxt(X) = Z Z . . .

(4) The C*-algebra C

(

A C) of the universal covering

A C of the group
A C of the ane transformations of the complex straight line can be included in
an extension
0 C(S
1
) /(H) C

(

A C) C(S
2
) 0
and its structure is uniquely dened by the invariant
Index C

(

A C) = 1 KK(S
2
, S
1
) = Z.
(5) The C*-algebra C

(G
,
) can be included into two subsequent extensions
(
1
) 0 C

(V
,
, F) C

(G
,
) C(S
1
) 0,
(
2
) 0 C(S
2
S
2
) /(H) C

(V
,
, F) C(S
1
) /(H) 0,
and its structure is uniquely dened by the KK-theory invariant
Index C

(G
,
) = ([
1
], [
2
]),
where
[
1
] KK(C(S
1
), C

(V
,
, F)),
172 Diep
[
2
] KK(S
1
, S
2
S
2
)

= Z Z
(6) The group C*-algebra C

(R
J
H
3
) can be included into two subsequent
extensions
(
1
) 0 C

(V
J
, F) C

(R
J
H
3
) C(S
1
) /(H) 0,
(
2
) 0 C

((R

R)
cpt
) /(H) C

(V
J
, F) C(S
1
) /(H) 0,
and its structure is uniquely dened by the KK-theory invariant
Index C

(R
J
H
3
) = ([
1
], [
2
]),
where
[
1
] KK(C(S
1
), C

(V
J
, F)),
[
2
] = (1, 1) KK(S
1
, S
1
)

= Z Z
(7) Finally, the C*-algebra C

(R H
3
) of the real diamond group can be
included in three subsequent extensions
(
1
) 0 C(S
2
S
2
) /(H) C

(R H
3
) A
1
0,
(
2
) 0 C(S
2
S
2
S
2
S
2
) /(H) A
1
A
2
0,
(
3
) 0 C
4
/(H) A
2
C(S
1
) 0,
and the structure of C

(R H
3
) is uniquely dened by the KK-theory invariant
Index C

(R H
3
) = ([
1
], [
2
], [
3
]),
where
[
1
] = (1, 1) KK(A
1
, C(S
2
S
2
)) = Z Z,
[
2
] =
_
_
_
_
_
1 0 0 1
1 1 0 0
0 1 1 0
0 0 1 1
_
_
_
_
_
KK(A
2
, C(S
2
S
2
S
2
S
2
))

= Hom
Z
(Z
4
, Z
4
),
[
3
] = (1, 1, 1, 1) KK(C(S
1
), C
4
) = Hom
Z
(Z, Z
4
) = Z
4
5. Deformation Quantization and Cyclic Theories
Let us nish this survey with some indication about some relations of the
problem with some new developments. Recall that the group algebra of nite or
compact groups are in fact some Hopf bialgebras. One deforms this Hopf bialgebra
structure to obtain the corresponding quantum groups. Our problem is therefore
closely related to the interesting problem to describe these quantum groups. One
of the methods is deformation quantization which is closely related to the orbit
method. The others which are closely related with KK-theory are the periodic
cyclic (co-)homologies. We nish this survey by indicating the subjects and the
author who is working in this area.
Diep 173
5.1. Star-Products and Star-Representations
See [Gu] and the references there.
5.2. Periodic Cyclic Homology
See [Cu], [CQ1] - [CQ3].
5.3. Chern Characters
See for example [Cu], [CQ3], [Pu].
References
[BDF] Brown, L.G., R.G. Douglas and P.A. Filmore, Extensions of C*-algebras
and K-homology, Ann. Math. (2) 105 (1977), 265324.
[BM] Boyer, R. and R. Martin, The group C*-algebra of the De Sitter group.
[C1] Connes, A., A survey of foliations and operator algebras, Proc. Symp. Pure
Math., 38(1982), 521628.
[C2] Connes, A., C*-alg`ebres et geometrie dierentielle, C. R. Acad. Sci. Paris,
Serie A, 280 (1980), 599604.
[Cu] J. Cuntz, J., A survey of some aspects of noncommutative geometry, Preprint,
Math. Inst. Uni Heidelberg, No 35(1992).
[CQ1] Cuntz, J. and D. Quillen, Algebra extensions and nonsingularity, Preprint,
Math. Inst. Uni Heidelberg, No 34(1992).
[CQ2] Cuntz, J. and D. Quillen, Operators on noncommutative dierential forms
and cyclic homology, Preprint, Math. Inst. Uni Heidelberg, No 38(1992).
[CQ3] Cuntz, J. and D. Quillen, Preprint in preparation, 1993.
[De] Delaroche, C., Extensions des C*-alg`ebres, Bull. Soc. Math. France, 29(1972).
[D1] Diep, D. N., The structure of the group C*-algebra of the group of ane
transformations of the straight line, Funkt. Anal. i Priloz., 9(1975), No 1, 63
64.
[D2] Diep, D. N., The structure of C*-algebras of type I, Vestnik Moskov. Uni.,
1978 , No 2, 81 87.
[D3] Diep, D. N., Applications of the homological K-functor cxt to studying the
structure of the C*-algebras of some solvable Lie groups, Ph. D. Thesis,
Moskov. Uni. 1977.
[D4] Diep, D. N.,Quelques aspects topologiques en analyse harmonique, Preprint
IHES/M/83/22; Acta Math. Vietnam., 8(1983), No 2, 35131.
174 Diep
[D5] Diep, D. N., Construction des representations unitaires par les K-orbites et
quantication, C. R. Acad. Paris, Serie A, 291(1980), 295298.
[D6] Diep, D. N.,Quantication des syst`emes hamiltoniens ` a laction plate dun
groupe de Lie, C. R. Acad. Sci. Paris, Serie I, 295(1982), 345348.
[D7] Diep, D. N., Ideaux de type compact associes aux representations irreductibles
induites par des repesentations liminaires de sous-groupes invariants, C. R.
Acad. Sci. Paris, Serie I, 294(1982), 189192.
[D8] Diep, D. N., Construction and reduction of the K-theory invariant Index C
(G) of group C*-algebras, Preprint SFB 343, Uni Bielefeld, 92-015 (1992).
[D9] Diep, D. N., Discrete series for loop groups I, Preprint SFB 343, Uni Bielefeld
92-015(1992).
[D10] Diep, D. N., Multidimensional quantization and Fourier integral operators,
Preprint, Math, Inst. Uni Heidelberg, No 52(1992).
[Do1] Dong, Tr. D., On globalization over U(1)-covering of Zuckermann (g, K)-
modules, Tap Chi Toan Hoc (J. of Math.), 19(1991), No 1, 60-72.
[Do2] Dong, Tr. D., A reduction of the globalization and U(1)-covering, Preprint
ICTP, Trieste, Italy, 1993.
[Du] Duo, M., Theorie de Mackey pour les groupes de Lie algebriques, Acta Math.
149(1982), 153213.
[DV] Dong, Tr. D. and Tr. Vui, On the procedure of multidimensional quantization,
Acta Math. Vietnam., 14(1989), 1930.
[E] Evans, B., C*-bundles and compact transformation groups, Memoirs A.M.S.,
No 269(1982).
[F] Fell, J. M. G., The structure of algebras of operator elds, Acta Math.,
106(1961), 233280.
[GN] Gelfand, I.M. and M.A. Naimark,Unitary representations of the group of
ane transformations of the straight line, Dokl. Akad. Nauk SSSR, 55(1947),
No 7, 571574.
[Gr] Green, P., C*-algebras of transformation groups with smooth orbit space,
Pacic J. Math., 72(1977), No 1.
[Gu] Gutt, S., Deformation quantization, Lectures at the Workshop on Represen-
tation Theory of Lie Groups, International Centre for Theoretical Physics,
Trieste, Italy, March 15 - April 2, 1993.
[JT] Jensen, K. and K. Thomsen, Elements of KK-theory, Birkh auser, Boston-
Basel-Berlin, 1991.
[K] Karoubi, M., K-theory: An introduction, Grundlehren der Math. Wis-
senschaften, No 226 , Springer-Verlag, Berlin-Heidelberg-New York, 1978.
Diep 175
[K1] Kasparov, G. G., The operator K- functor and extensions of C*-algebras,
Math. USSR Izvestija, 16(1981), No 3, 513572.
[K2] Kasparov, G.G., The operator K-theory and its applications, Itogi Nauki
Tekh., Ser. Soverem. Probl. Math., 27(1985), 331.
[Ki] Kirillov, A. A., Elements of the theory of representations, Springer-Verlag,
Berlin-Heidelberg-New York, 1976.
[KM] Kraljevic, H. and D. Milicic, The C*-algebra of the universal covering group
of SL(2, R), Glasnik Math., Ser. III, 1972, 3548.
[M1] Milicic, D., Topological representations of the group C*-algebra of SL(2, R),
Glasnik Math., Ser. III, 1971,231246.
[M2] Milicic, D., Algebraic D-modules and representations of semisimple Lie
groups, Lectures at the Workshop on Representation Theory of Lie Groups,
International Centre for Theoretical Physics, Trieste, Italy, March 15 - April
2, 1993.
[P] Perdrizet, F., Topologie et trace sur les C*-alg`ebres, Bull. Soc. Math. France,
99(1971), 193239.
[Pu] Puschnigg, M., Asymptotic cyclic cohomology, Dissertation, 1993, Univ. Hei-
delberg.
[R1] Rosenberg, J., The C*-algebras of some real and p-adic solvable groups,
Pacic J. Math., 65(1976), No 1, 175192.
[R2] Rosenberg, J., K-theory of group C*-algebras, foliation C*-algebras and
crossed products, Contemp. Math., 1988, 251301.
[RS] Rosenberg, J. and C. Schochet, The classication of extensions of C*-
algebras, Bull. A.M.S., New series, 4(1981), 105-110.
[SW] Sulanke, R. and P. Wintgen, Dierentialgeometrie und Faserb undel, Berlin,
Verl. Deutscher Verlag der Wissenschaften.
[SV] Son, V. M., and H. H. Viet, Sur la structure des C*-alg`ebres dune classe de
groupes de Lie, J. Operator Theory, 11(1984),7790.
[T] Torpe, A. M., K-theory for the leaf space of foliations by Reeb components, J.
Funct. Anal., 61(1985), 1571.
[V1] Viet, H. H., Sur la structure des C*-alg`ebres dune classe de groupes de Lie
resolubles de dimension 3, Acta Math. Vietnam., 11(1986), No 1, 8691.
[V2] Viet, H. H., Application of the sequence of repeated extensions to studying
the structure of type I C*-algebras, Preprint Series, Inst. Math. Hanoi, No
87.24(1987), (in Russian).
[Vo] Voiculescu, D., Remarks on the singular extensions in C*-algebra of the
Heisenberg group, J. Operator Theory.
176 Diep
[Vu1] Vu, L. A., On the structure of the C*-algebra of the foliation formed by
the K-orbits of maximal dimension of the real diamond group, J. Operator
Theory, 24(1990), No 2, 227238.
[Vu2] Vu, L. A., On the foliation formed by the generic K-orbits of the MD4-
groups, Acta Math. Vietnam., 15(1990), No 2, 3555.
[Vui1] Vui, Tr., Multidimensional quantization and U(1)-covering, Acta Math.
Vietnam., 16(1991), 103119.
[Vui2] Vui, Tr., A reduction of the multidimensional quantization and U(1)-
covering, Tap Chi Toan Hoc (J. Math.), 19(1992), 112.
[Vui3] Vui, Tr., Geometric construction of unitary representations of Lie groups
via multidimensional quantization, Preprint ICTP, Trieste, Italy, 1993.
[W1] Wang, X., Les C*-alg`ebres dune classe de groupes de Lie resolubles, C. R.
Acad. Sci. Paris, Serie I, 306(1988), 765767.
[W2] Wang, X., On the C*-algebras of a family of solvable Lie groups and folia-
tions, Ph. D. Thesis, Univ. Maryland, 1985.
[Wa] Wasserman, A., Une demonstration de la conjecture de Connes-Kasparov
pour les groupes de Lie lineaires connexes reductifs, Preprint IHES, 1983.
[Wo] Wolf, J. A., Admissible representations and the geometry of ag manifold,
Lectures at the Workshop on Representation Theory of Lie Groups, Inter-
national Centre for Theoretical Physics, Trieste, Italy, March 15 - April 2,
1993; to appear in Proceedings of the A.M.S. Summer Conference Penrose
Transform and Analytic Cohomology.
Mathematisches Institut der Univer-
sit at Heidelberg
Im Neuenheimer Feld 288
6900 Heidelberg, Germany
[email protected]
Received May 28, 1993