You are on page 1of 94

ANINTRODUCTIONTO

ANALYSIS ONFRACTALS
ALEXANDER TEPLYAEV
UNIVERSITY OF CONNECTICUT
HTTP://WWW.MATH.UCONN.EDU/~ TEPLYAEV/GREGYNOG/
LMS/EPSRC SHORT COURSE
10 - 15 JANUARY 2007
GREGYNOG HALL
IN CONJUNCTION WITH
ANALYSIS ON GRAPHS AND ITS APPLICATIONS
ISAAC NEWTON INSTITUTE FOR MATHEMATICAL SCIENCES
CAMBRIDGE, UK, 8 JANUARY - 29 JUNE 2007
Major general references (books)
[1] M. T. Barlow, Diusions on fractals. Lectures on Probability Theory and
Statistics (Saint-Flour, 1995), 1121, Lecture Notes in Math., 1690, Springer,
Berlin, 1998.
[2] J. Kigami, Analysis on fractals. Cambridge Tracts in Mathematics 143, Cam-
bridge University Press, 2001.
[3] M. L. Lapidus and M. van Frankenhuysen, Fractal geometry, complex dimen-
sions and zeta functions. Geometry and spectra of fractal strings. Springer
Monographs in Mathematics. Springer, New York, 2006.
[4] R. S. Strichartz, Dierential equations on fractals: a tutorial. Princeton
University Press, 2006.
Contents
Major general references (books) 2
Lecture 1
Laplacians on self-similar graphs
and relation to self-similar groups 6
Abbreviated list of references 30
Lecture 2
Laplacians on self-similar fractals
and spectral zeta functions 36
Abbreviated list of references 54
Lecture 3
Kigamis resistance forms on fractals
and relation to quantum graphs 64
Abbreviated list of references 87
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T
0
v v
v v v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
v v v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T
0
Lecture 1
Laplacians on self-similar graphs
and relation to self-similar groups
R. Rammal and G. Toulouse, Random walks on fractal structures and percolation
clusters. J. Physique Letters 44 (1983), L13L22.
R. Rammal, Spectrum of harmonic excitations on fractals. J. Physique 45 (1984).
S. Alexander, Some properties of the spectrum of the Sierpi nski gasket in a mag-
netic eld. Phys. Rev. B 29 (1984).
E. Domany, S. Alexander, D. Bensimon and L. Kadano, Solutions to the Schrodinger
equation on some fractal lattices. Phys. Rev. B (3) 28 (1984).
Y. Gefen, A. Aharony and B. B. Mandelbrot, Phase transitions on fractals.
I. Quasilinear lattices. II. Sierpi nski gaskets. III. Innitely ramied lattices. J. Phys.
A 16 (19831984).
R. B. Stinchcombe, Fractals, phase transitions and criticality. Fractals in the natural
sciences. Proc. Roy. Soc. London Ser. A 423 (1989), 1733.
J. Bellissard, Renormalization group analysis and quasicrystals, Ideas and methods
in quantum and statistical physics (Oslo, 1988). Cambridge Univ. Press, 1992.

2
1
1
1
z =
3
2

1
2
1
2
1
2
1
2
1
2
1
2
2 1
1
1 1
z =
3
4

2
1
1
1
1
1
1
z =
3
2



1 1
1
1
1
1
z =
5
4
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
v v v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
v v v
v
v v
v
v v
v v v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T
0
Let be the probabilistic Laplacian (generator of a simple random walk) on the
Sierpi nski lattice. If z ,=
3
2
,
5
4
,
1
2
, and R(z) = z(4z + 5), then
R(z) () z ()
() = J
R

D
where D
def
= |
3
2
|

m=0
R
m
|
3
4
|
_
and J
R
is the Julia set of R(z).
E
T
0
3
2
5
4
3
4
1
2
3
2
5
4
,
, , ,
,
, , , , ,
There are uncountably many nonisomorphic
Sierpi nski lattices.
Theorem (T). The spectrum of is pure point.
Eigenfunctions with nite support are complete.
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v
v v
vv v
v
v v
v

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T
0

r
r
r
r
r
r
r
r
r
r
r
r
r
r
r
r
r
r
r
r
r
r
r
r
r
r
r
r
r
r

_ _ _
_ _ _ _ _ _
`
_
`
_
`
_
`
_
`
_
`
_
Let
(0)
be the Laplacian with zero (Dirichlet) boundary condition at L. Then
the compactly supported eigenfunctions of
(0)
are not complete (eigenvalues in
E is not the whole spectrum).

_
0
0
0
1
g
1g
0
0 0
0
0 0
0
0 0

. . .

. . .
Let L
(0)
be the set of two points adjacent to L and
(0)

be the spectral
measure of
(0)
associated with 1
L
(0)
. Then supp(
(0)

) = J
R
has Lebesgue
measure zero and
d(
(0)

R
1,2
)
d
(0)

(z) =
(8z + 5)(2z + 3)
(2z + 1)(4z + 5)
Fix p, q>0, p+q=1, and dene probabilistic Laplacians
n
on the segments
[0, 3
n
] of Z
+
inductively as a generator of the random walks:
0 1
g g
E '
1 1
0 1 3
g g g g
E E E ' ' '
1 q p p q 1
0 1 3 6 9
g g g g g g g g g g
E E E E E E E E E ' ' ' ' ' ' ' ' '
1 q p p q q p q p p q p q q p p q 1
0 3
n
2(3
n
) 3
n+1
g , , , , , , , , , g g g
E E E ' ' '
1 q p p q 1
and let = lim
n

n
be the corresponding probabilistic Laplacian on Z
+
.
If z ,=1p and R(z)=z(z
2
+3z+2+pq)/pq, then R(z) (
n
)
z (
n+1
)
E
T
0
2
2
1 p
, , ,
Theorem(T). () = J
R
, the Julia set of R(z).
If p=q, then ()=[2, 0], spectrum is a.c.
If p ,= q, then () is a Cantor set of Lebesgue measure zero, spectrum is
singularly continuous.
There are uncountably many random self-similar Laplacians on Z:
For a sequence K = |k
j
|

j=1
, k
j
|0, 1, 2|, let
X
n
=
n

j=1
k
j
3
j
and
n
is a probabilistic Laplacian on [X
n
, X
n
+3
n
]:
X
n
X
n
+3
n1
X
n
+2(3
n1
) X
n
+3
n
g , , , , , , , , , g g g
E E E ' ' '
1 q p p q 1
In the previous example k
j
= 0 for all j.
Theorem (T).
For any sequence K we have () = J
R
. The same is true for the Dirichlet
Laplacian on Z
+
(when k
j
0).
R. Grigorchuk and Z. Sunik, Asymptotic aspects of Schreier graphs and Hanoi
Towers groups, preprint.
,
,

, ,

, ,

,
,

, ,

,
,

, ,

,
,

, ,

, ,

,
,

,,

,
,

,
,

,,

,
,

,
,

,,

,
,

,
,

,,

,
,

,
,

,,

,
,

,
,

,,

,
,

,
,

,,

,
,

,
,

,,

,
,

,
,

,,

,
,

,
,

,,

,
,

,
,

,,

,
,

,
,

,,

,
,

,
,

,,

,
,

, ,

,
,

, ,

,
,

, ,

,
,

,,

,
,

,
,

,,

,
,

,
,

,,

,
,

,
,

,,

,
,

,
,

,,

,
,

,
,

,,

,
,

,
,

,,

,
,

,
,

,,

,
,

,
,

,,

,
Sierpi nski 3-graph Sierpi nski 4-graph
(Hanoi Towers-3 group) (standard)
These three polynomials are conjugate:
Sierpi nski 3-graph (Hanoi Towers-3 group): f(x) = x
2
x 3
f(3) = 3, f
t
(3) = 5
Sierpi nski 4-graph, adjacency matrix Laplacian: P() = 5
2
P(0) = 0, P
t
(0) = 5
Sierpi nski 4-graph, probabilistic Laplacian: R(z) = 4z
2
+ 5z
R(0) = 0, R
t
(0) = 5
,

, ,

,
,

, ,

,
,

, ,

,
,

,,

,
,

,
,

,,

,
,

,
,

,,

,
,

,
,

,,

,
,

,
,

,,

,
,

,
,

,,

,
,

,
,

,,

,
,

,
,

,,

,
,

,
,

,,

,
,

, ,

,
,

, ,

,
,

, ,

,
,

,,

,
,

,
,

,,

,
,

,
,

,,

,
,

,
,

,,

,
,

,
,

,,

,
,

,
,

,,

,
,

,
,

,,

,
,

,
,

,,

,
,

,
,

,,

,
,
,
,
,
,
, ,
,
,
,
,
,
,
,
, ,
,
,
,
,
,
,
,
, ,
,
,
,

,
,

,,

,
,

,
,

,,

,
,

,
,

,,

,
,

,
,

,,

,
,

,
,

,,

,
,

,
,

,,

,
,

,
,

,,

,
,

,
,

,,

,
,

,
,

,,

,
,
,
, , ,
,
, ,
,
,
, , ,
,
, , ,
,
, ,
,
,
, ,
,
,
, , ,
,
, ,
,
,
, , ,
,
, , ,
,
, ,
,
,
, , ,
,
, , ,
,
, ,
,
,
, ,
,
,
, , ,
,
, ,
,
,
, ,
,
,
, , ,
,
, ,
,
,
, , ,
,
, , ,
,
, ,
,
,
, ,
,
,
, , ,
,
, ,
,
,
, ,
,

, ,

,
,

, ,

,
,

, ,

,
,
, , ,
,
, ,
,
,
, , ,
,
, , ,
,
, ,
,
,
, ,
,
,
, , ,
,
, ,
,
,
, , ,
,
, , ,
,
, ,
,
,
, , ,
,
, , ,
,
, ,
,
,
, ,
,
,
, , ,
,
, ,
,
,
, ,
,
,
, , ,
,
, ,
,
,
, , ,
,
, , ,
,
, ,
,
,
, ,
,
,
, , ,
,
, ,
,
,
, ,
,
,
, , ,
,
, ,
,
,
, , ,
,
, , ,
,
, ,
,
,
, ,
,
,
, , ,
,
, ,
,
,
, , ,
,
, , ,
,
, ,
,
,
, , ,
,
, , ,
,
, ,
,
,
, ,
,
,
, , ,
,
, ,
,
,
, ,
,
,
, , ,
,
, ,
,
,
, , ,
,
, , ,
,
, ,
,
,
, ,
,
,
, , ,
,
, ,
,
,
, ,
,

,
,

,,

,
,

,
,

,,

,
,

,
,

,,

,
,

,
,

,,

,
,

,
,

,,

,
,

,
,

,,

,
,

,
,

,,

,
,

,
,

,,

,
,

,
,

,,

,
,
,
, , ,
,
, ,
,
,
, , ,
,
, , ,
,
, ,
,
,
, ,
,
,
, , ,
,
, ,
,
,
, , ,
,
, , ,
,
, ,
,
,
, , ,
,
, , ,
,
, ,
,
,
, ,
,
,
, , ,
,
, ,
,
,
, ,
,
,
, , ,
,
, ,
,
,
, , ,
,
, , ,
,
, ,
,
,
, ,
,
,
, , ,
,
, ,
,
,
, ,
Theorem. Eigenvalues and eigenfunctions on the
Sierpi nski 3-graphs and Sierpi nski 4-graphs are in
one-to-one correspondence, with the exception of the
eigenvalue z =
3
2
for the 4-graphs.
,

,

4z
2
+ 5z
,

,
,

,,

,
,

,

4
3
z
2
+
8
3
z
,
,
, ,
, ,

2z
2
+ 4z

, ,
,

, ,

,
,

, ,

,
,

, ,

,
,
, , ,
,
, ,
,
,
, , ,
,
, , ,
,
, ,
,
,
, ,
,
,
, , ,
,
, ,
,
,
, , ,
,
, , ,
,
, ,
,
,
, , ,
,
, , ,
,
, ,
,
,
, ,
,
,
, , ,
,
, ,
,
,
, ,
,
,
, , ,
,
, ,
,
,
, , ,
,
, , ,
,
, ,
,
,
, ,
,
,
, , ,
,
, ,
,
,
, ,
,

,
,

,,

,
,

,
,

,,

,
,

,
,

,,

,
,

,
,

,,

,
,

,
,

,,

,
,

,
,

,,

,
,

,
,

,,

,
,

,
,

,,

,
,

,
,

,,

,
Sierpi nski 3-graph Sierpi nski 4-graph
(Hanoi Towers-3 group) (standard)
R(z) = 2z
2
+ 4z R(z) =
4
3
z
2
+
8
3
z
Let H and H
0
be Hilbert spaces, and U : H
0
H be an isometry.
Denition. We call an operator H spectrally similar to an operator H
0
with
functions
0
and
1
if
U

(H z)
1
U = (
0
(z)H
0

1
(z))
1
In particular, if
0
(z) ,= 0 and R(z) =
1
(z)/
0
(z), then
U

(H z)
1
U =
1

0
(z)
(H R(z))
1
.
If H =
_
S

X
X Q
_
then
S zI
0


X(QzI
1
)
1
X =
0
(z)H
0

1
(z)I
0
Theorem (Malozemov, Teplyaev). If is the graph Laplacian on a self-
similar symmetric innite graph, then
J
R
(

) J
R
D

where D

is a discrete set and J


R
is the Julia set of the rational function R.
Abbreviated list of references
[1] S. Alexander, Some properties of the spectrum of the Sierpi nski gasket in
a magnetic eld. Phys. Rev. B 29 (1984), 5504-5508.
[2] Bartholdi, Laurent ; W. Woess, Spectral computations on lamplighter groups
and Diestel-Leader graphs. J. Fourier Anal. Appl. 11 (2005), 175202.
[3] J. Bellissard, Renormalization group analysis and quasicrystals, Ideas and
methods in quantum and statistical physics (Oslo, 1988), 118148. Cambridge
Univ. Press, Cambridge, 1992.
[4] S. Broerio, W. Woess, Green kernel estimates and the full Martin bound-
ary for random walks on lamplighter groups and Diestel-Leader graphs.
Ann. Inst. H. Poincare Probab. Statist. 41 (2005), 11011123.
[5] S. Broerio, W. Woess, Positive harmonic functions for semi-isotropic ran-
dom walks on trees, lamplighter groups, and DL-graphs. Potential Anal. 24
(2006), 245265.
[6] E. Domany, S. Alexander, D. Bensimon and L. Kadano, Solutions to the
Schrodinger equation on some fractal lattices. Phys. Rev. B (3) 28 (1984),
31103123.
[7] Y. Gefen, A. Aharony and B. B. Mandelbrot, Phase transitions on fractals. I.
Quasilinear lattices. II. Sierpi nski gaskets. III. Innitely ramied lattices.
J. Phys. A 16 (1983), 12671278; 17 (1984), 435444 and 12771289.
[8] S. Goldstein, Random walks and diusions on fractals, in Percolation The-
ory and Ergodic Theory of Innite Particle Systems (H.Kesten, ed.), 121129,
IMA Math. Appl., Vol. 8, Springer, New York, 1987.
[9] P. Grabner, Functional iterations and stopping times for Brownian motion
on the Sierpi nski gasket. Mathematika 44 (1997), 374400.
[10] R. Grigorchuk and Z. Sunik, Asymptotic aspects of Schreier graphs and
Hanoi Towers groups, preprint.
[11] V. A. Kamanovich and A. M. Vershik, Random walks on discrete groups:
boundary and entropy. Ann. Probab., 11 (1983), 457490.
[12] B. Kron, Green functions on self-similar graphs and bounds for the spec-
trum of the Laplacian. Ann. Inst. Fourier (Grenoble) 52 (2002), 18751900.
[13] B. Kron, Growth of self-similar graphs. J. Graph Theory 45 (2004), 224239.
[14] B. Kron and E. Teu, Asymptotics of the transition probabilities of the
simple random walk on self-similar graph, Trans. Amer. Math. Soc., 356
(2003) 393414.
[15] T. Lindstrm, Brownian motion on nested fractals. Mem. Amer. Math. Soc.
420, 1989.
[16] L. Malozemov and A. Teplyaev, Pure point spectrum of the Laplacians on
fractal graphs. J. Funct. Anal. 129 (1995), 390405.
[17] L. Malozemov and A. Teplyaev, Self-similarity, operators and dynamics.
Math. Phys. Anal. Geom. 6 (2003), 201218.
[18] B. B. Mandelbrot, Hasards et tourbillons: quatre contes `a clef. Annales des
Mines, November 1975, 6166.
[19] B. B. Mandelbrot, Les Objets Fractals. Forme, Hasard et Dimension. Nou-
velle Biblioth`eque Scientique. Flammarion, Editeur, Paris, 1975.
[20] B. B. Mandelbrot, Fractals: form, chance, and dimension. Translated from
the French. Revised edition. W. H. Freeman and Co., San Francisco, Calif.,
1977.
[21] B. B. Mandelbrot, The Fractal Geometry of Nature. W. H. Freeman and
Co., San Francisco, Calif., 1982.
[22] G. Musser, Wireless communications, Scientic American, July 1999.
[23] C. Puente-Baliarda, J. Romeu., Ra. Pous and A. Cardama, On the behavior
of the Sierpinski multiband fractal antenna. IEEE Trans. Antennas and
Propagation 46 (1998), 517524.
[24] R. Rammal, Spectrum of harmonic excitations on fractals. J. Physique 45
(1984), 191206.
[25] R. Rammal and G. Toulouse, Random walks on fractal structures and per-
colation clusters. J. Physique Letters 44 (1983), L13L22.
[26] C. Sabot, Pure point spectrum for the Laplacian on unbounded nested
fractals. J. Funct. Anal. 173 (2000), 497524.
[27] C. Sabot, Integrated density of states of self-similar Sturm-Liouville opera-
tors and holomorphic dynamics in higher dimension. Ann. Inst. H. Poincare
Probab. Statist. 37 (2001), 275311.
[28] L. Rogers, R. Strichartz and A. Teplyaev, Smooth bumps, a Borel theorem
and partitions of unity on p.c.f. fractals preprint.
[29] C. Sabot, Existence and uniqueness of diusions on nitely ramied self-
similar fractals. Ann. Sci.

Ecole Norm. Sup. (4) 30 (1997), 605673.
[30] C. Sabot, Electrical networks, symplectic reductions, and application to
the renormalization map of self-similar lattices. J. Physique Letters 44
(1983), L13L22. Fractal Geometry and Applications: A Jubilee of Benoit
Mandelbrot, Part 1. Proceedings of Symposia in Pure Mathematics 72, Amer.
Math. Soc., (2004), 155205.
[31] T. Shima, On eigenvalue problems for Laplacians on p.c.f. self-similar
sets, Japan J. Indust. Appl. Math. 13 (1996), 123.
[32] R. B. Stinchcombe, Fractals, phase transitions and criticality. Fractals in
the natural sciences. Proc. Roy. Soc. London Ser. A 423 (1989), 1733.
[33] R. S. Strichartz, The Laplacian on the Sierpinski gasket via the method of
averages. Pacic J. Math. 201 (2001), 241256.
[34] A. Teplyaev, Spectral Analysis on Innite Sierpi nski Gaskets, J. Funct.
Anal., 159 (1998), 537-567.
[35] E. Teu, The average displacement of the simple random walk on the
Sierpi nski graph. Combin. Probab. Comput. 12 (2003), 203222.
[36] W. Woess, A note on the norms of transition operators on lamplighter
graphs and groups. Internat. J. Algebra Comput. 15 (2005), 12611272.
[37] W. Woess, Lamplighters, Diestel-Leader graphs, random walks, and har-
monic functions. Combin. Probab. Comput. 14 (2005), 415433.
[38] W. Woess, Random walks on innite graphs and groups. Cambridge Tracts
in Mathematics, 138. Cambridge University Press, Cambridge, 2000.
Lecture 2
Laplacians on self-similar fractals
and spectral zeta functions
Three contractions F
1
, F
2
, F
3
: R
1
R
1
, F
j
(x) =
1
3
(x+p
j
), with xed
points p
j
= 0,
1
2
, 1. The interval I=[0, 1] is a unique compact set such that
I =
_
j=1, 2, 3
F
j
(I)
The boundary of I is I = V
0
= |0, 1| and the discrete approxima-
tions to I are V
n
=

j=1, 2, 3
F
j
(V
n1
) =
_
k
3
n
_
3
n
k=0
V
0
=I : _ _
c

q
V
1
: _ _ _ _
c

q
V
2
: _ _ _ _ _ _ _ _ _ _
Denition. The discrete Dirichlet (energy) form on V
n
is
E
n
(f) =

x,yV
n
yx
(f(y)f(x))
2
and the Dirichlet (energy) form on I is E(f) = lim
n
3
n
E
n
(f) =
_
1
0
|f
t
(x)|
2
dx
Denition. A function h is harmonic if it minimizes the energy given the
boundary values.
Proposition. 3E
n+1
(f) E
n
(f) and 3E
n+1
(h) = E
n
(h) = 3
n
E(h)
for a harmonic h.
Proposition. The Dirichlet (energy) form on I is self-similar in the sense that
E(f) = 3

j = 1, 2, 3
E(fF
j
)
Denition. The discrete Laplacians on V
n
are

n
f(x) =
1
2

yV
n
yx
f(y)f(x), xV
n
\V
0
and the Laplacian on I is f(x) = lim
n
9
n

n
f(x) = f
tt
(x)
GaussGreen (integration by parts) formula:
E(f) =
_
1
0
ffdx +ff
t

1
0
Spectral asymptotics: Let () be the eigenvalue counting function of
the Dirichlet or Neumann Laplacian :
() = #|j :
j
< |.
Then
lim

()

d
s
/2
=
1

where d
s
= 1 is the spectral dimension.
Denition. The spectral zeta function is

(s) =

j
,=0
_

j
_
s/2
Its poles are the complex spectral dimensions.
Let R(z) be a polynomial of degree N such that its Julia set J
R
(, 0],
R(0) = 0 and c = R
t
(0) > 1.
Denition. The zeta function of R(z) for Re(s) > d
R
=
2 log N
log c
is

z
0
R
(s) = lim
n

zR
n
|z
0
|
(c
n
z)
s/2
=

s/2
j
Theorem.
z
0
R
(s) =
f
1
(s)
1Nc
s/2
+f
z
0
2
(s), where f
1
(s) and f
z
0
2
(s) are ana-
lytic for Re(s)>0. If J
R
is totally disconnected, then this meromorphic continuation
extends to Re(s)>, where >0.
In the case of polynomials this theorem has been improved by Grabner et al.
d
R
the poles of
z
0
R

_
2 log N+4in
log c
: nZ
_
0
E

d
R

"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
Theorem.

(s) =
0
R
(s) where R(z) = z(4z
2
+12z+9).
The Riemann zeta function (s) satises (s) =
s

0
R
(s) The only complex
spectral dimension is the pole at s = 1.
A sketch of the proof: If z,=
1
2
,
3
2
, then
R(z) (
n
) z (
n+1
)
and so

(s) =
0
R
(s) since the eigenvalues
j
of are limits of the eigenvalues
of 9
n

n
.
Also
j
=
2
j
2
and so

(s) =

j=1
_

2
j
2
_
s/2
=
s
(s)
where (s) is the Riemann zeta function. Q.E.D.
(s) =
s
lim
n

zR
n
|0|
z,=0
_
9
n
z
_
s/2
Denition.

is Laplacian if
E(f) =
_
1
0
|f
t
(x)|
2
dx=
_
1
0
f

fd +ff
t

1
0
.
Denition. A probability measure is self-similar with weights m
1
, m
2
, m
3
if =

j=1, 2, 3
m
j
F
j
.
Proposition.

f(x)=
f
tt

= lim
n
_
1+
2
pq
_
n

n
f(x).

n
f(
k
3
n
)=
_
pf(
k1
3
n
) +qf(
k+1
3
n
) f(
k
3
n
)
qf(
k1
3
n
) +pf(
k+1
3
n
) f(
k
3
n
)
where m
1
=m
3
, p=
m
2
m
1
+m
2
, q=
m
1
m
1
+m
2
, and
_ _ _ _
E E E ' ' '
1
m
1
m
2
m
3
q p p q 1
_ _ _ _ _ _ _ _ _ _
E E E E E E E E E ' ' ' ' ' ' ' ' '
1 q p p q q p q p p q p q q p p q 1
Spectral asymptotics: If () is the eigenvalue counting function of the Dirichlet
or Neumann Laplacian

, then
0 < liminf

()

d
s
/2
limsup

()

d
s
/2
<
where the spectral dimension is
d
s
=
log 9
log(1+
2
pq
)
1.
All the inequalities are strict if and only if p ,= q.
Proposition. R(z) (
n
) z (
n+1
)
where z,=1p and R(z)=z(z
2
+3z+2+pq)/pq.
Note that R
t
(0)=1 +
2
pq
, and d
s
=d
R
.
Theorem.

(s)=
0
R
(s)
Three contractions F
1
, F
2
, F
3
: R
2
R
2
,
F
j
(x) =
1
2
(x+p
j
), with xed points p
1
, p
2
, p
3
.
p
1
p
2
p
3
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
.
. .
The Sierpi nski gasket is a unique compact set S such that
S =

j=1, 2, 3
F
j
(S)
Denition. The boundary of S is
S = V
0
= |p
1
, p
2
, p
3
|
and discrete approximations to S are
V
n
=

j=1, 2, 3
F
j
(V
n1
)
V
0
:
y y
y

T
T
T
T
T
T
T
T
T
T
V
1
:
x x
xx x
x
x x
x

T
T
T
T
T

T
T
T
T
T

T
T
T
T
T
V
2
:
w w
ww w
w
w w
w
w w
ww w
w
w w
w
w w
ww w
w
w w
w

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T

T
T
T
Denition. The discrete Dirichlet (energy) form on V
n
is
E
n
(f) =

x,yV
n
yx
(f(y)f(x))
2
and the Dirichlet (energy) form on S is
E(f) = lim
n
_
5
3
_
n
E
n
(f)
Denition. A function h is harmonic if it minimizes the energy given the
boundary values.
Proposition.
5
3
E
n+1
(f) E
n
(f)
5
3
E
n+1
(h)=E
n
(h)=
_
5
3
_
n
E(h) for a harmonic h.
Theorem (Kigami). E is a local regular Dirichlet form on S which is self-similar
in the sense that
E(f) =
5
3

j = 1, 2, 3
E(fF
j
)
Denition. The discrete Laplacians on V
n
are

n
f(x) =
1
4

yV
n
yx
f(y)f(x), xV
n
\V
0
and the Laplacian on S is

f(x) = lim
n
5
n

n
f(x)
if this limit exists and

f is continuous.
GaussGreen (integration by parts) formula:
E(f) =
_
S
f

fd +

pS
f(p)
n
f(p)
where is the normalized Hausdor measure, which is self-similar with weights
1
3
,
1
3
,
1
3
:
=
1
3

j = 1, 2, 3
F
j
.
Spectral asymptotics: If () is the eigenvalue counting function of the Dirichlet
or Neumann Laplacian

, then
0 < liminf

()

d
s
/2
< limsup

()

d
s
/2
<
where the spectral dimension is
1 < d
s
=
log 9
log 5
< 2.
Proposition. R(z) (
n
) z (
n+1
) where z,=
1
2
,
3
4
,
5
4
and R(z) = z(5 + 4z).
Theorem (Fukushima, Shima). Every eigenvalue of

has a form
=5
m
lim
n
5
n
R
n
(z
0
)
where R
n
(z
0
) is a preimage of z
0
=
3
4
,
5
4
under the n-th iteration power
of the polynomial R(z). The multiplicity of such an eigenvalue is C
1
3
m
+C
2
.
Theorem. Zeta function of the Laplacian on the Sierpi nski gasket is

(s) =
1
2

3
4
R
(s)
_
1
5
s/2
3
+
3
5
s/2
1
_
+
1
2

5
4
R
(s)
_
35
s/2
5
s/2
3

5
s/2
5
s/2
1
_
d
s
d
R
=
log 9
log 5
=
log 4
log 5
T
0
d
R
1 d
s
_
_
_
_
_
_
_
g
g
g
g
g
g
g
_
_
_
_
_
_
_
E
1
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
"
Denition. If L is a fractal string, that is, a disjoint collection of intervals of
lengths l
j
, then its geometric zeta function is
L
(s) =

l
s
j
.
Theorem (Lapidus). If A=
d
2
dx
2
is a Neumann or Dirichlet Laplacian on L,
then
A
(s) =
s
(s)
L
(s).
Example: Cantor self-similar fractal string.
If L is the complement of the middle third Cantor set in [0, 1], then the complex
spectral dimensions are 1 and |
log 2+2in
log 3
: nZ|,

L
(s) =
1
123
s
,
A
(s) = (s)

s
123
s
E
T
0
log 2
log 3
1
_
_
_
_
_
_
_
_
Denition. A post critically nite (p.c.f.) self-similar set F is a compact con-
nected metric space with a nite boundary F F and contractive injections

i
: F F such that
F = (F) =
k
_
i=1

i
(F)
and

v
(F)

w
(F)
v
(F)

w
(F),
for any two dierent words v and w of the same length. Here for a nite word
w |1, . . . , k|
m
we dene
w
=
w
1
. . .
w
m
.
We assume that F is a minimal such subset of F. We call
w
(F) an m-cell.
The p.c.f. assumption is that every boundary point is contained
in a single 1-cell.
Theorem (Kigami, Lapidus). The spectral dimension of the Laplacian

is
the unique solution of the equation
k

i=1
(r
i

i
)
d
s
/2
= 1
Conjecture. On every p.c.f. fractal F there exists a local regular Dirichlet form E
which gives positive capacity to the boundary points and is self-similar in the sense
that
E(f) =
k

i=1

i
E(f
i
)
for a set of positive renement weights = |
i
|
k
i=1
.
Denition. The group G of acts on a nitely ramied fractal F if each g G is
a homeomorphism of F such that g(V
n
) = V
n
for all n 0.
Proposition. Suppose a group G of acts on a self-similar nitely ramied fractal
F and G restricted to V
0
is the whole permutation group of V
0
. Then there exists
a unique, up to a constant, G-invariant self-similar resistance form E with equal
energy renormalization weights
i
and
E
0
(f, f) =

x,yV
0
_
f(x) f(y)
_
2
.
Moreover, for any G-invariant self-similar measure the Laplacian

has the
spectral self-similarity property (a.k.a. spectral decimation).
Abbreviated list of references
[1] B. Adams, S.A. Smith, R. Strichartz and A. Teplyaev, The spectrum of
the Laplacian on the pentagasket. Fractals in Graz 2001, Trends Math.,
Birkhauser (2003).
[2] M. T. Barlow, Diusions on fractals. Lectures on Probability Theory and
Statistics (Saint-Flour, 1995), 1121, Lecture Notes in Math., 1690, Springer,
Berlin, 1998.
[3] M. T. Barlow and R. F. Bass, Brownian motion and harmonic analysis on
Sierpinski carpets. Canad. J. Math., 51 (1999), 673744.
[4] M. T. Barlow and R. F. Bass, Random walks on graphical Sierpinski carpets.
Random walks and discrete potential theory (Cortona, 1997), 2655, Sympos.
Math., XXXIX, Cambridge Univ. Press, Cambridge, 1999.
[5] M. T. Barlow and R. F. Bass, Stability of parabolic Harnack inequalities.
Trans. Amer. Math. Soc., 356 (2004), 15011533.
[6] M. T. Barlow, R. F. Bass and T. Kumagai, Stability of parabolic Harnack
inequalities on metric measure spaces. J. Math. Soc. Japan 58 (2006) 485-
519.
[7] M. T. Barlow and B. M. Hambly, Transition density estimates for Brownian
motion on scale irregular Sierpinski gaskets. Ann. Inst. H. Poincare Probab.
Statist., 33 (1997), 531557.
[8] M. T. Barlow and J. Kigami, Localized eigenfunctions of the Laplacian on
p.c.f. self-similar sets. J. London Math. Soc., 56 (1997), 320332.
[9] M.T. Barlow and E.A. Perkins, Brownian motion on the Sierpinski gasket.
Probab. Theory Related Fields 79 (1988), 543623.
[10] O. Ben-Bassat, R. S. Strichartz and A. Teplyaev, What is not in the domain
of the Laplacian on Sierpinski gasket type fractals. J. Funct. Anal. 166
(1999), 197217.
[11] N. Ben-Gal, A. Shaw-Krauss, R. S. Strichartz, C. Young, Calculus on the
Sierpinski gasket. II. Point singularities, eigenfunctions, and normal
derivatives of the heat kernel. Trans. Amer. Math. Soc. 358 (2006), 3883
3936.
[12] E.J. Bird, S.-M. Ngai and A. Teplyaev, Fractal Laplacians on the Unit In-
terval, Ann. Sci. Math. Quebec 27 (2003), 135168.
[13] K. Coletta, K. Dias, R. S. Strichartz, Numerical analysis on the Sierpinski
gasket, with applications to Schrodinger equations, wave equation, and
Gibbs phenomenon. Fractals 12 (2004), 413449.
[14] K. Dalrymple, R. S. Strichartz and J. P. Vinson, Fractal dierential equations
on the Sierpinski gasket. J. Fourier Anal. Appl., 5 (1999), 203284.
[15] M. Denker and S. Koch, Hausdor dimension for Martin metrics. Algebraic
and topological dynamics, 163170, Contemp. Math., 385, Amer. Math. Soc.,
2005.
[16] M. Denker and H. Sato, Reections on harmonic analysis of the Sierpi nski
gasket. Math. Nachr., 241 (2002), 3255.
[17] G. Derfel, P. Grabner and F. Vogl, The zeta function of the Laplacian on
certain fractals, preprint (2005).
[18] M. Fukushima, Dirichlet forms, diusion processes and spectral dimen-
sions for nested fractals. Ideas and methods in Mathematical Analysis,
Stochastics, and applications (Oslo, 1988), 151161, Cambridge Univ. Press,
Cambridge, 1992.
[19] M. Fukushima and T. Shima, On a spectral analysis for the Sierpi nski
gasket. Potential Analysis 1 (1992), 1-35.
[20] M. Gibbons, A. Raj and R. S. Strichartz, The nite element method on the
Sierpinski gasket. Constr. Approx. 17 (2001), 561588.
[21] P. Grabner and W. Woess (editors) Fractals in Graz 2001. Analysis
dynamicsgeometrystochastics. Proceedings of the conference held at
Graz University of Technology, Graz, June 2001. Edited by . Trends in Math-
ematics. Birkhauser Verlag, Basel, 2003.
[22] B. M. Hambly, Heat kernels and spectral asymptotics for some random
Sierpinski gaskets. Fractal geometry and stochastics, II (Greifswald/Koserow,
1998), 239267, Progr. Probab., 46, Birkhauser, Basel, 2000.
[23] B. M. Hambly, On the asymptotics of the eigenvalue counting function for
random recursive Sierpinski gaskets. Probab. Theory Related Fields, 117
(2000), 221247.
[24] W. Hansen and M. Zahle, Restricting isotropic -stable Levy processes
from R
n
to fractal sets. Forum Math. 18, 171191.
[25] P. E. Herman, R. Peirone, R. S. Strichartz, p-energy and p-harmonic func-
tions on Sierpinski gasket type fractals. Potential Anal. 20 (2004), 125148.
[26] R. G. Hohlfeld and N. Cohen, SelfSimilarity and the Geometric Require-
ments for Frequency Independence in Antennae. Fractals 7 (1999), 7984.
[27] J. Kigami, A harmonic calculus on the Sierpi nski spaces. Japan J. Appl.
Math. 6 (1989), 259290.
[28] J. Kigami, Harmonic calculus on p.c.f. selfsimilar sets. Trans. Amer.
Math. Soc. 335 (1993), 721755.
[29] J. Kigami, Analysis on fractals. Cambridge Tracts in Mathematics 143,
Cambridge University Press, 2001.
[30] J. Kigami and M. L. Lapidus, Weyls problem for the spectral distribution of
Laplacians on p.c.f. self-similar fractals. Comm. Math. Phys. 158 (1993),
93125.
[31] J. Kigami and M. L. Lapidus, Selfsimilarity of volume measures for Lapla-
cians on p.c.f. selfsimilar fractals. Comm. Math. Phys. 217 (2001), 165
180.
[32] S. Kozlov, Harmonization and homogenization on fractals. Comm. Math.
Phys. 153 (1993), 339357.
[33] S. Kusuoka, Lecture on diusion process on nested fractals. Lecture Notes
in Math. 1567 3998, Springer-Verlag, Berlin, 1993.
[34] S. Kusuoka and X. Y. Zhou, Dirichlet forms on fractals: Poincare constant
and resistance. Probab. Theory Related Fields 93 (1992), 169196.
[35] M. L. Lapidus, Spectral and fractal geometry: from the Weyl-Berry conjec-
ture for the vibrations of fractal drums to the Riemann zeta-function. Dif-
ferential equations and mathematical physics (Birmingham, AL, 1990), 151
181, Math. Sci. Engrg., 186, Academic Press, Boston, MA, 1992.
[36] M. L. Lapidus and M. van Frankenhuysen, Fractal Geometry and Number
Theory. Complex Dimensions of Fractal Strings and Zeros of Zeta Func-
tions. Birkhauser, Boston, 2000.
[37] M. L. Lapidus and M. van Frankenhuysen, Fractality, self-similarity and
complex dimensions. Fractal Geometry and Applications: A Jubilee of Benoit
Mandelbrot, Part 1. Proceedings of Symposia in Pure Mathematics 72, Amer.
Math. Soc., (2004), 349372.
[38] M. L. Lapidus and M. van Frankenhuysen, Fractal geometry, complex dimen-
sions and zeta functions. Geometry and spectra of fractal strings. Springer
Monographs in Mathematics. Springer, New York, 2006.
[39] M. L. Lapidus and H. Maier, Hypoth`ese de Riemann, cordes fractales vi-
brantes et conjecture de Weyl-Berry modiee. Acad. Sci. Paris Ser. I Math.
313, (1991), 1924.
[40] M. L. Lapidus and H. Maier, The Riemann hypothesis and inverse spectral
problems for fractal strings. J. London Math. Soc. (2) 52, (1995), 1534.
[41] M. L. Lapidus and C. Pomerance, Fonction zeta de Riemann et conjecture
de Weyl-Berry pour les tambours fractals. C. R. Acad. Sci. Paris Ser. I
Math. 310, (1990), 343348.
[42] M. L. Lapidus and C. Pomerance, The Riemann zeta-function and the one-
dimensional Weyl-Berry conjecture for fractal drums. Proc. London Math.
Soc. (3) 66, (1993), 4169.
[43] T. Lindstrm, Brownian motion on nested fractals. Mem. Amer. Math. Soc.
420, 1989.
[44] R. S. Strichartz, Piecewise linear wavelets on Sierpinski gasket type frac-
tals. J. Four. Anal. Appl., 3 (1997), 387-416.
[45] R. S. Strichartz, Some properties of Laplacians on fractals. J. Funct. Anal.,
164 (1999), 181-208.
[46] R. S. Strichartz, Analysis on fractals. Notices AMS, 46 (1999), 11991208.
[47] R. S. Strichartz, Isoperimetric estimates on Sierpinski gasket type fractals.
Trans. Amer. Math. Soc. 351 (1999), 17051752.
[48] R. S. Strichartz, Taylor approximations on Sierpinski type fractals. J.
Funct. Anal. 174 (2000), 76127.
[49] R. S. Strichartz, Evaluating integrals using self-similarity. Amer. Math.
Monthly 107 (2000), 316326.
[50] R. S. Strichartz, Function spaces on fractals. J. Funct. Anal. 198 (2003),
4383.
[51] R. S. Strichartz, Fractafolds based on the Sierpinski and their spectra.
Trams. AMS 355 (2003), 40194043.
[52] R. S. Strichartz, Solvability for dierential equations on fractals. J. Anal.
Math. 96 (2005), 247267.
[53] R. S. Strichartz, Laplacians on fractals with spectral gaps have nicer
Fourier series. Math. Res. Lett. 12 (2005), 269274.
[54] R. S. Strichartz, Analysis on products of fractals. Trans. Amer. Math. Soc.
357 (2005), 571615.
[55] R. S. Strichartz, Dierential equations on fractals: a tutorial. Princeton
University Press, 2006.
[56] R. S. Strichartz and M. Usher, Splines on fractals. Math. Proc. Cambridge
Philos. Soc. 129 (2000), 331360.
[57] R. S. Strichartz, C. Wong, The p-Laplacian on the Sierpinski gasket. Non-
linearity 17 (2004), 595616.
[58] A. Teplyaev, Spectral Analysis on Innite Sierpi nski Gaskets, J. Funct.
Anal., 159 (1998), 537-567.
[59] A. Teplyaev, Spectral zeta functions of fractals and the complex dynamics
of polynomials, to appear in the Transactions of the American Mathematical
Society.
[60] M. Zahle, Harmonic calculus on fractalsa measure geometric approach.
II. Trans. Amer. Math. Soc. 357 (2005), 34073423.
Lecture 3
Kigamis resistance forms on fractals
and relation to quantum graphs
Denition. A compact connected metric space F is called a nitely ramied
self-similar set if there are injective contraction maps
1
, ...,
m
: F F
and a nite set V
0
F such that
F =
m
_
i=1

i
(F) = (F)
and
F
w
F
w
t = V
w
V
w
t
for any two distinct words w, w
t
W
n
= |1, ..., m|
n
, where F
w
=
w
(F),
V
w
=
w
(V
0
) and
w
=
w
1
...
w
n
.
The vertices of generation n are dened by V
n
= (V
n1
) =
n
(V
0
).
The fractal F can be uniquely reconstructed from its combinatorial skeleton or
ancestor: |F = V
0
, V
1
, |
V
0
| [Kigami, 1993, Appendix A].
A symmetric vanishing on the diagonal function c
0
: V
2
0
R
+
(set of conduc-
tances) denes a discrete Dirichlet form
E
0
(f) =

x,yV
0
(f(y) f(x))
2
c
0
(x, y).
Its renement by is
E
1
(f) =

(E
0
)(f) =
k

i=1

i
E
0
(f
i
).
and the trace map is
Tr(E
1
)(f) = inf |E
1
(g)|g : V
1
R, g|
V
0
= f|.
Theorem (Kigami). For given a set of positive renement weights = |
i
|
k
i=1
self-similar local regular Dirichlet forms E which gives positive capacity to the
boundary points are in one-to-one correspondence with the xed points E
0
of the
renormalization map

= Tr

.
Denition. A resistance form E is self-similar if
E(f, f) =
m

i=1

i
E(f
i
, f
i
).
Conjecture. Any nitely ramied self-similar set has a self-similar resistance form.
Any p.c.f. self-similar set has a regular self-similar resistance form.
Thus we are looking for nonlinear eigenvectors E
0
D P

(E
0
) = E
0
where D is the cone of Dirichlet forms on V
0
with and P is the cone of nonnegative
quadratic forms. Its interior P

consists of positive forms.


Proposition.
(1)

: D D, P P, P

.
(2)

is continuous on D P

(3)

(E) =

(E) for all 0


(4)

(E +F)

(E) +

(F)
Hilberts projective metric (a pseudo distance on P

) is
h(E/F) = ln
M(E/F)
m(E/F)
.
where E,FP

is the biggest lower bound of E/F,


m(E/F) = sup| > 0|F E| > 0
and M(E/F) = m(F/E)
1
.
Proposition.
(1) h(E, F) = h(E, F) for all , > 0.
(2) Let H=|EB|trace(E)=1| (an ane hyperplane). Then (H P

, h)
is a complete metric space.
(3) The h- and the ] ]-topology coincide on H P

.
(4) h(E, F) = 0 if and only if E = F.
(5) lim
FP
h(E, F) = +
(6)

is h-nonexpansive on P

, that is, lower q

-level sets are

-invariant.
Let H = H D P

and q

: H R
+
,
q

(E) = h(

(E), E).
Proposition.
(1)

has a unique eigenvector F H if and only if q

|
H
vanishes only at F.
(2)

has multiple eigenvectors in H if and only if q

vanishes on a connected
set which accumulates at P.
(3) When a

-forward orbit started in H is contained in B


r
(E) for some r > 0
and E H, then there exists a

-eigenvector in B
3r
(E) H.
Proposition. Let |
n
| be such that

n
converges to in (C(H), ] ]

). If
q = h((), ) : H R
+
vanishes only at a single point, then there exists an
m N such that

n
has a unique eigenvector in H, for n m.
Denition. A collection of renement weights is admissible if and only if

(E
0
) = E
0
has a solution E
0
D P

.
Proposition. The set of admissible weights is open.
Theorem. (Hambly, Metz, T.) Let
n
,

(0, ]
k
and

has a
unique eigenvector in H. Then there exist nite admissible renement weights.
This result can be summarized as follows: If, by collapsing a subset of cells of F,
one can obtain a structure which has admissible weights, then F also has admissible
nite weights.
Proposition. If #V
0
= 3 then admissible weights exist.
Denition. The group G of acts on a nitely ramied fractal F if each g G is
a homeomorphism of F such that g(V
n
) = V
n
for all n 0.
Proposition. Suppose a group G of acts on a self-similar nitely ramied fractal
F and G restricted to V
0
is the whole permutation group of V
0
. Then there exists
a unique, up to a constant, G-invariant self-similar resistance form E with equal
energy renormalization weights
i
and E
0
(f, f) =

x,yV
0
_
f(x) f(y)
_
2
.
Theorem (Hambly, Metz, T.) Suppose a self-similar nitely ramied fractal
F has connected interior and a group G acts on F such that its action on V
0
is
transitive. Then there exists a G-invariant self-similar resistance form E on F.
Theorem (Hambly, Metz, T.) Suppose a self-similar nitely ramied fractal F
has connected interior and a symmetric boundary. Then there exists a G-invariant
self-similar resistance form E on F.
Examples.
Generalized non-symmetric Sierpi nski gaskets in R
2
:

1
2
+
1
3
>
1
1

1
1
+
1
2
>
1
3

1
1
+
1
3
>
1
2
Cut Sierpi nski gasket:

1
+
2
= 1

3
+
2
= 1
Unit interval:

1
+
2
= 1
Vicsek set:

1
+
3
+
5
= 1

2
+
4
+
5
= 1
A generalized Vicsek set
A generalized Sierpi nski gasket
GRAPH-DIRECTED FRACTALS
v
1
v
2
v
3
v
4
v
5
v
6
v
7
v
8
v
9
v
10
v
11
v
12
The house fractal.
Denition. A pair (E, DomE) is a resistance form on a countable set V

if
DomE is a linear subspace of (V

) containing constants, E is a nonnegative


symmetric quadratic form on DomE, and E(u, u) = 0 if and only if u is
constant.
Let be an equivalence relation on DomE dened by u v if and only if
u v is constant on V

. Then (E/, DomE) is a Hilbert space.


For any nite subset V V

and for any v (V ) there exists


u DomE such that u

V
= v.
For any p, q V

there exists the eective resistance between metric


R(p, q) = sup
_
_
u(p)u(q)
_
2
E(u,u)
: uDomE
_
<
Hence any u DomE has a unique R-Holder continuous extension to ,
the R-completion of V

.
Markov property: for any u DomE we have that E( u, u) E(u, u),
where
u(p) =
_

_
1 if u(p) 1,
u(p) if 0 < u(p) < 1,
0 if u(p) 1.
For any nite subset U V

the nite dimensional Dirichlet form E


U
on U is
E
U
(f, f) = inf |E(g, g) : g DomE, g

U
= f|
and is called the trace of E on U.
If U
1
U
2
then E
U
1
is the trace of E
U
2
on U
1
.
Theorem (Kigami). Suppose that V
n
are nite subsets of V

and that

n=0
V
n
is R-dense in V

. Then
E(f, f) = lim
n
E
V
n
(f, f)
for any f DomE, where the limit is non-decreasing.
Theorem (Kigami). Suppose that V
n
are nite sets, and the nite dimensional
resistance forms E
V
n
on V
n
are compatible: each E
V
n
is the trace of E
V
n+1
on V
n
.
Then there exists a resistance form E on V

n=0
V
n
such that
E(f, f) = lim
n
E
V
n
(f, f)
for any f DomE, and the limit is non-decreasing.
Denition. A nitely ramied fractal F is a compact metric space with
a cell structure F = |F

|
A
and a boundary (vertex) structure
V = |V

|
A
such that the following conditions hold.
A is a countable index set;
each F

is a distinct compact connected subsets of F;


each V

is a nite subset of F

with at least two elements;


if F

k
j=1
F

j
then V



k
j=1
V

j
;
there exists a ltration |A
n
|

n=0
such that
(1) A
n
are nite subsets of A, A
0
= |0|, F
0
= F;
(2) A
n
A
m
= if n ,= m;
(3) for any A
n
there are
1
, ...,
k
A
n+1
such that F

k
j=1
F

j
;
F

= V

for any two distinct ,


t
A
n
;
for any strictly decreasing innite sequence of cells there exists x F such
that

n1
F

n
= |x|.
If these conditions are satised, then
(F, F, V) = (F, |F

|
A
, |V

|
A
)
is called a nitely ramied cell structure.
Denition. A function is harmonic if it minimizes the energy for the given set of
boundary values. A function is n-harmonic if it minimizes the energy for the given
set of values on V
n
.
Theorem. Suppose that all n-harmonic functions are continuous. Then any con-
tinuous function is R-continuous, and any R-Cauchy sequence converges in the
topology of F. Also, there is a continuous injective map : F which is the
identity on V

.
Then we can (and will) consider as a subset of F. Then is the R-closure of
V

. In a sense, is the set where the Dirichlet form E lives.


Theorem. Suppose that all n-harmonic functions are continuous. Then E is a
local regular Dirichlet form on (with respect to any measure that charges every
nonempty open set).
Denition. We x a complete, up to constant functions, energy orthonormal set
of harmonic functions h
1
, ..., h
k
, where k = |V
0
| 1, and dene the Kusuoka
energy measure by
=
h
1
+... +
h
k
.
If F

t F

, then
M
,
t : (V

) (V

t)
is the linear map which is dened as follows. If f

is a function on V

then let h
f

be the unique harmonic function on F

that coincides with f

on V

. Then we
dene
M
,
tf

= h
f

t
.
Proposition. If F

j
then D

,
j
D

j
M
,
j
and
(F

) = Tr M

where M

= M
0,
and D

is the matrix of the Dirichlet form E

on V

.
We denote Z

=
M

(F

)
if (F

) ,= 0. Then we dene matrix valued


functions Z
n
(x) = Z

if (F

) ,= 0, A
n
and x F

\V

. Note that
Tr Z
n
(x) = 1 by denition.
Theorem. For -almost all x there is a limit Z(x) = lim
n
Z
n
(x).
Proof. One can see, following Kusuokas idea, that Z
n
is a bounded -martingale.

The energy measures


h
are the same as the energy measures in the general theory
of Dirichlet forms. The matrix Z is the matrix whose entries are the densities
Z
ij
=
d
h
i
,h
j
d
It has been recently proved by Hino that is singular with respect to any product
measure for a large class of fractals.
Theorem. If the space of piecewise harmonic functions is dense in DomE then
any f DomE has a weak gradient f such that
E(f, f) =
_
F
{f, Zf)d
Conjecture. For any nitely ramied fractal
rankZ(x) = 1
for -almost all x.
This has been recently proved by Hino for a large class of p.c.f. fractals.
GRADIENT IN HARMONIC COORDINATED
Let V
0
= |v
1
, ..., v
m
| and let h
j
be the unique harmonic function with boundary
values h
j
(v
i
) =
i,j
.
Kigamis harmonic coordinate map : F R
m
is
(x)=(h
1
(x), ..., h
m
(x)).
In what follows we assume that : F F
H
= (F) is a homeomorphism,
F = F
H
, (x) = x and identify (V
0
) with R
m
in the natural way.
Theorem. If f is the restriction to F of a C
1
(R
m
) function then f DomE,
and such functions are dense in DomE. Moreover,
E(f, f) =
_
F
{f, Zf)d
for any f C
1
(R
m
).
We have the analog of the Gauss-Green formula:
E(f, g) =
_
F
g

fd,
for any function g DomE, vanishing on the boundary V
0
, and any function
f Dom

, where

is the energy Laplacian.


Theorem. If f is the restriction to F of a C
2
(R
m
) function then f Dom

,
and such functions are dense in Dom

. Moreover, -almost everywhere

f = Tr (ZD
2
f)
where D
2
f is the matrix of the second derivatives of f.
Conjecture. On the Sierpi nski gasket, if f Dom

then f is the restriction


to F of a C
1
(R
m
) function.
We also can dene a dierent sequence of approximating energy forms. In various
situations these forms are associated with so called quantum graphs, photonic
crystals and cable systems. If f C
1
(R
m
) then
E
Q
n
(f, g) =

x,yV
n
c
n,x,y
E
Q
x,y
(f, f)
where
E
Q
x,y
(f, f) =
_
1
0
_
d
dt
f
_
x(1 t) +ty
_
_
2
dt
is the integral of the square of the derivative
d
dt
f
_
x(1 t) +ty
_
= {f
_
x(1 t) +ty
_
, y x)
of f along the straight line segment connecting x and y. Thus E
Q
x,y
(f, f) is the
usual one dimensional energy of a function on a straight line segment.
If f is linear then E
Q
x,y
(f, f) =
_
f(x) f(y)
_
2
. Therefore if f is piecewise
harmonic then E
Q
n
(f, f) = E
n
(f, f) for all large enough n.
Therefore for any C
1
(R
m
)-function we have
lim
n
E
Q
n
(f, f) = E(f, f)
It is easy to see that if g is a C
1
(R
m
)-function vanishing on V
0
and f is a
C
2
(R
m
)-function then
E
Q
n
(f, g) =

x,yV
n
c
n,x,y
_
1
0
g
_
x(1 t) +ty
_
_
d
2
dt
2
f
_
x(1 t) +ty
_
_
dt
because after integration by parts all the boundary terms are canceled. Then if
A
n
then

x,yV

c
n,x,y
d
2
dt
2
f
_
x(1 t) +ty
_
=

x,yV

c
n,x,y
m

i,j=1
D
2
ij
f
_
x(1 t) +ty
_
(y
i
x
i
)(y
j
x
j
) =
Tr
_
M

_
D
2
f(x

) +R
n
(x, y, t, f, , x

)
_
_
where x

and
lim
n
|R
n
(x, y, t, f, , x

)| = 0
uniformly.
Let H
x
be the space of harmonic functions on F that vanishes at x.
Denition. If h H
x
then the intrinsic derivative
df
dh
(x) R exists if
f(y) = f(x) +h(y)
df
dh
(x) +o

h(y)

yx
.
The intrinsic gradient Grad
x
f H
x
exists if for any non constant h H
x
f(y) = f(x) + Grad
x
(y) +o

h(y)

yx
.
Theorem (Pelander, T). Let be a self-similar measure on a p.c.f. s-s set with
weights
j
. Let
+
and

be the upper and lower Lyapunov exponents of the


matrices M
j
with respect to the measure and log =

m
j=1

j
log(r
j

j
).
If
+
> then
df
dh
(x) exists for any f Dom

, any non constant h H


and -almost all x.
If

> then Grad


x
f exists for any f Dom

, any non constant h H


and -almost all x.
Abbreviated list of references
[1] N. Bouleau and F. Hirsch, Dirichlet forms and analysis on Wiener space.
de Gruyter Studies in Math. 14, 1991.
[2] B. Boyle, D. Ferrone, N. Rifkin, K. Savage and A. Teplyaev, Electrical Re-
sistance of N-gasket Fractal Networks, to appear in the Pacic Journal of
Mathematics.
[3] P. Doyle and J.L. Snell, Random walks and electric networks. Carus Math-
ematical Monographs, 22, MAA, 1984.
[4] E.B. Dynkin and A.A. Yushkevich, Markov processes: Theorems and prob-
lems. Translated from the Russian. Plenum Press, New York 1969.
[5] D. Fontaine, T. Smith and A. Teplyaev, Random Sierpi nski gasket, Quan-
tum Graphs and Their Applications, Contemporary Mathematics 415 (2006),
AMS, Providence, RI.
[6] D. Fontaine and A. Teplyaev, Greens function and eigenfunctions on ran-
dom Sierpi nski gaskets, preprint.
[7] M. Fukushima, Y. Oshima and M. Takada, Dirichlet forms and symmetric
Markov processes. deGruyter Studies in Math. 19, 1994.
[8] B. M. Hambly, V. Metz and A. Teplyaev, Admissible renements of energy
on nitely ramied fractals, J. London Math. Soc. 74 (2006), 93112.
[9] M. Hino, On singularity of energy measures on self-similar sets. Probab.
Theory Related Fields 132 (2005), 265290.
[10] M. Hino and T. Kumagai, A trace theorem for Dirichlet forms on fractals.
J. Funct. Anal. 238 (2006), 578611.
[11] M. Hino and K. Nakahara, On singularity of energy measures on self-
similar sets II, preprint.
[12] J. Kigami, Harmonic metric and Dirichlet form on the Sierpi nski gasket.
Asymptotic problems in probability theory: stochastic models and diusions
on fractals (Sanda/Kyoto, 1990), 201218, Pitman Res. Notes Math. Ser., 283,
Longman Sci. Tech., Harlow, 1993.
[13] J. Kigami, Eective resistances for harmonic structures on p.c.f. self-
similar sets. Math. Proc. Cambridge Philos. Soc. 115 (1994), 291303.
[14] J. Kigami, Analysis on fractals. Cambridge Tracts in Mathematics 143,
Cambridge University Press, 2001.
[15] J. Kigami, Harmonic analysis for resistance forms. J. Functional Analysis
204 (2003), 399444.
[16] J. Kigami, Local Nash inequality and inhomogeneity of heat kernels. Proc.
London Math. Soc. (3) 89 (2004), 525544.
[17] J. Kigami, D. R. Sheldon and R. S. Strichartz, Greens functions on fractals.
Fractals 8 (2000), 385402.
[18] J. Kigami, R. S. Strichartz, K. C. Walker, Constructing a Laplacian on the
diamond fractal. Experiment. Math. 10 (2001), 437448.
[19] P. Kuchment, Quantum graphs I. Some basic structures. Waves in random
media, 14 (2004), S107S128.
[20] P. Kuchment, Quantum graphs II. Some spectral properties of quantum
and combinatorial graphs. J. Phys. A. 38 (2005), 48874900.
[21] S. Kusuoka, Dirichlet forms on fractals and products of random matrices.
Publ. Res. Inst. Math. Sci. 25 (1989), 659680.
[22] R. D. Mauldin and M. Urba nski, Dimension and measures for a curvilinear
Sierpinski gasket or Apollonian packing. Adv. Math., 136 (1998), 2638.
[23] V. Metz, How many diusions exist on the Vicsek snowake? Acta Appl.
Math. 32 (1993), 227241.
[24] V. Metz, The cone of diusions on nitely ramied fractals. Nonlinear
Anal. 55 (2003), 723738.
[25] V. Metz, Renormalization contracts on nested fractals. J. Reine Angew.
Math. 480 (1996), 161175.
[26] V. Metz, Shorted operators: an application in potential theory. Linear Al-
gebra Appl. 264 (1997), 439455.
[27] V. Metz and K.-T. Sturm, Gaussian and non-Gaussian estimates for heat
kernels on the Sierpi nski gasket. Dirichlet forms and stochastic processes
(Beijing, 1993), 283289, de Gruyter, Berlin, 1995.
[28] R. Meyers, R. Strichartz and A. Teplyaev, Dirichlet forms on the Sierpinski
gasket. Pacic J. Math. 217 (2004), 149174.
[29] A.

Oberg, R. S. Strichartz and A.Q. Yingst, Level sets of harmonic functions
on the Sierpinski gasket. Ark. Mat. 40 (2002), 335362.
[30] K. A. Okoudjou, R. S. Strichartz, Weak uncertainty principles on fractals.
J. Fourier Anal. Appl. 11 (2005), 315331.
[31] K. Okoudjou, L. Salo-Coste and A. Teplyaev, Innite dimensional i.f.s. and
smooth functions on the Sierpinski gasket, to appear in the Transactions of
the American Mathematical Society.
[32] R. Peirone, Convergence and uniqueness problems for Dirichlet forms on
fractals. Boll. Unione Mat. Ital. Sez. B Artic. Ric. Mat. (8) 3 (2000), 431460.
[33] R. Peirone, Convergence of discrete Dirichlet forms to continuous Dirich-
let forms on fractals. Potential Anal. 21 (2004), 289309.
[34] A. Pelander and A. Teplyaev, Innite dimensional i.f.s. and smooth func-
tions on the Sierpinski gasket, to appear in the Indiana Journal of Mathe-
matics (available online).
[35] A. Pelander and A. Teplyaev, Products of random matrices and derivatives
on p.c.f. fractals, preprint.
[36] J. Stanley, R. Strichartz and A. Teplyaev, Energy partition on fractals. In-
diana Univ. Math. J. 52 (2003), 133156.
[37] R. S. Strichartz, Harmonic mappings of the Sierpinski gasket to the circle.
Proc. Amer. Math. Soc. 130 (2002), 805817.
[38] A. Teplyaev, Gradients on fractals. J. Funct. Anal. 174 (2000) 128154.
[39] A. Teplyaev, Energy and Laplacian on the Sierpi nski gasket. Fractal Geom-
etry and Applications: A Jubilee of Benoit Mandelbrot, Part 1. Proc. Sympos.
Pure Math. 72, Amer. Math. Soc., (2004), 131154.
[40] A. Teplyaev, Harmonic coordinates on fractals with nitely ramied cell
structure, to appear in the Canadian Journal of Mathematics.
[41] E. Teu, On the Hausdor dimension of the Sierpi nski gasket with re-
spect to the harmonic metric. Fractals in Graz 2001, 263269, Trends Math.,
Birkhauser, Basel, 2003.
Sierpi nski gasket
in harmonic coordinates