Quantum Graphs PDF

An introduction to Quantum Graphs theory
Mario Arioli
STFC Rutherford Appleton Laboratory
Toulouse 9
th
December 2010
Quantum Graphs, Toulouse 9
th
December 2010 Mario Arioli, RAL
Overview of talk
Motivations and Modelling
Graphs, metric graphs, and quantum graphs
Self-adjoint Hamiltonians and boundary conditions
Modelling again
Waves and eigenvalues problems
Numerical issues and domain decomposition
Open problems i.e.

2 / 42
th
Overview of talk
Motivations and Modelling
Graphs, metric graphs, and quantum graphs
Self-adjoint Hamiltonians and boundary conditions
Modelling again
Waves and eigenvalues problems
Numerical issues and domain decomposition
Open problems i.e. The things I have not understood yet!!!

2 / 42
th
Example: Naphthalene
3 / 42
th
Example: Polystyrene
4 / 42
th
Example: Graphene
5 / 42
th
Example: spectral clustering
6 / 42
th
Example: Human body
7 / 42
th
Modelling (examples)
A fat graph (l
e
length of edge e)
Dicult to have a decent

triangulation of the fat domain!
Irregular solution (corners)
Given an Hamiltonian on the fat

graph, to what does it converge
when 0?
8 / 42
th
A fat graph (l
e
length of edge e)


when 0?
8 / 42
th
A fat graph (l
e
length of edge e)


when 0?
8 / 42
th
A fat graph (l
e
length of edge e)


when 0?
8 / 42
th
Graph (e edge and v vertex)


when 0?
8 / 42
th
Combinatorial and metric Graphs
A Combinatorial Graph is dened by a set 1 = v

j
of
vertices and a set c = e
k
of edges connecting the vertices
that can be nite or countably innite. Each edge e can be
identied by the couple of vertices that it connects
(e = (v
j
1
, v
j
2
)).
A graph is a Metric Graph if at each edge e is assigned a

length l
e
(0, ) and a measure (normally the Lebesgue
one). Each edge can be assimilated to a nite or innite
segment of the real line (0, l
e
) R, with the natural
coordinate s
e
.
9 / 42
th

j
of
k
(e = (v
j
1
, v
j
2
)). NO GEOMETRY

length l
e
e
coordinate s
e
.
9 / 42
th

j
of
k
(e = (v
j
1
, v
j
2
)).

length l
e
e
coordinate s
e
.
9 / 42
th
Remarks
We can remove vertices of degree 2 after we fuse the 2 edges

into one
The graph is a topological manifold (or a 1D simplicial

complex) having singularities at the vertices, i.e. it is NOT a
dierentiable manifold.
is provided with a global metric and the distance between

two points (not necessarily vertices) is the length of the
shortest path between them. Thus, the points on are the
vertices and all the points on the edges. The Lebesgues
measure is well dened on all of for nite graphs.
is NOT necessarily embedded in a Euclidean space (R

n
)
10 / 42
th
Remarks

into one



n
)
10 / 42
th
Remarks

into one



n
)
10 / 42
th
Remarks

into one



n
)
10 / 42
th
Conditions on innite graphs
Condition A An edge of innite length has only one vertex. It is a
ray starting from a vertex.
Condition B For any positive number r and any vertex v there is
only a nite number of vertices w at a distance less
than r from v.
11 / 42
th
Conditions on innite graphs
Condition A An edge of innite length has only one vertex. It is a
ray starting from a vertex.
Condition B For any positive number r and any vertex v there is
only a nite number of vertices w at a distance less
than r from v.
11 / 42
th
Hilbert spaces
Denition-L
2
():
L
2
() =
eE
L
2
(e)
f (s) L
2
() i [[f [[
2
L
2
()
=
eE
[[f [[
2
L
2
(e)
<
Denition-H
1
() (Sobolev space) :
H
1
() =
_
eE
H
1
(e)
_
C
0
()
f (s) H
1
() i [[f [[
2
H
1
()
=
eE
[[f [[
2
H
1
(e)
<
C
0
() space of continuous functions on .
12 / 42
th
Quantum Graphs
Let 1 an operator (Hamiltonian) dened on H
1
().
A Quantum Graph is a metric graph where an
Hamiltonian 1 and boundary conditions that assure 1
is self-adjoint are dened.
13 / 42
th
Hamiltonian
Operators (s denotes the coordinate on an edge)
Second derivative f
d
2
f
ds
2
Schrodinger f
d
2
f
ds
2
+ V(s)f
Magnetic Schrodinger f
_
1
i
d
ds
A(s)
_
2
f + V(s)f
Others: pseudo-dierential, higher order derivative, etc...
A natural condition is to assume that f (e) H
2
(e), e c.
14 / 42
th
Hamiltonian
Second derivative f
d
2
f
ds
2
Schrodinger f
d
2
f
ds
2
+ V(s)f
_
1
i
d
ds
A(s)
_
2
f + V(s)f
2
(e), e c.
14 / 42
th
Hamiltonian
Second derivative f
d
2
f
ds
2
Schrodinger f
d
2
f
ds
2
+ V(s)f
_
1
i
d
ds
A(s)
_
2
f + V(s)f
2
(e), e c.
14 / 42
th
Hamiltonian
Second derivative f
d
2
f
ds
2
Schrodinger f
d
2
f
ds
2
+ V(s)f
_
1
i
d
ds
A(s)
_
2
f + V(s)f
2
(e), e c.
14 / 42
th
Hamiltonian
Second derivative We will focus on f
d
2
f
ds
2
Schrodinger f
d
2
f
ds
2
+ V(s)f
_
1
i
d
ds
A(s)
_
2
f + V(s)f
2
(e), e c.
14 / 42
th
Boundary conditions
We are interested in local conditions at the vertices. Let d be the
degree of vertex v. For functions f
j
H
2
(e
v
) on the edges
connected at v, we expect boundary conditions involving the
values of the functions and their directional derivative taken in the
outgoing directions at the vertex v:
A
v
F + B
v
F
= 0
where A R
dd
and B R
dd
, F = (f
1
(v), . . . , f
d
(v) and
F
= (f

1
(v), . . . , f

d
(v).
15 / 42
th
Boundary conditions
We are interested in local conditions at the vertices. Let d be the
degree of vertex v. For functions f
j
H
2
(e
v
) on the edges
connected at v, we expect boundary conditions involving the
values of the functions and their directional derivative taken in the
outgoing directions at the vertex v:
A
v
F + B
v
F
= 0
where A R
dd
and B R
dd
, F = (f
1
(v), . . . , f
d
(v) and
F
= (f

1
(v), . . . , f

d
(v).
The rank of the matrices [A
v
, B
v
] R
d2d
must be equal to d.
15 / 42
th
Finite graphs
Theorem
Let be a metric graph with nitely many edges. Consider the
operator 1 acting as
d
2
ds
2
on each edge e c, with the domain
consisting of the functions f H
2
(e) on e and satisfying the
conditions
A
v
F + B
v
F
= 0
at each vertex v 1.
Let
_
A
v
R
d
v
d
v
, B
v
R
d
v
d
v
[v 1
_
a collection of matrices
such that Rank(A
v
, B
v
) = d
v
for all v.
1 is self-adjoint i v 1, A
v
B
T
v
= B
v
A
T
v
(Kostrykin Schrader, 1999) (Kuchment, 2004)
16 / 42
th
A linear algebra bit (1)
We drop the subscript v for a moment
B = WV
T
= (W
1
, W
2
)
_

1
0
0 0
__
V
T
1
V
T
2
_
(SVD)
17 / 42
th
B = WV
T
= (W
1
, W
2
)
_

1
0
0 0
__
V
T
1
V
T
2
_
(SVD)
A = WRV
T
R =
_
R
11
R
12
R
21
R
22
_
17 / 42
th
B = WV
T
= (W
1
, W
2
)
_

1
0
0 0
__
V
T
1
V
T
2
_
(SVD)
A = WRV
T
R =
_
R
11
R
12
R
21
R
22
_
BA
T
= AB
T
= R
21
= 0
17 / 42
th
B = WV
T
= (W
1
, W
2
)
_

1
0
0 0
__
V
T
1
V
T
2
_
(SVD)
A = WRV
T
R =
_
R
11
R
12
R
21
R
22
_
BA
T
= AB
T
= R
21
= 0
W
T
(A, B)
_
V 0
0 V
_
=
_
R
11
R
12

1
0
0 R
22
0 0
_
17 / 42
th
W
T
(A, B)
_
V
T
F
V
T
F
_
= 0
_
R
11
R
12

1
0
0 R
22
0 0
_
_
_
_
_
V
T
1
F
V
T
2
F
V
T
1
F
V
T
2
F
_
_
_
_
= 0
18 / 42
th
W
T
(A, B)
_
V
T
F
V
T
F
_
= 0
_
R
11
R
12

1
0
0 R
22
0 0
_
_
_
_
_
V
T
1
F
V
T
2
F
V
T
1
F
V
T
2
F
_
_
_
_
= 0
Rank
_
(A, B)
_
= d = R
22
invertible = V
T
2
F = 0 = F span(V
1
)
18 / 42
th
W
T
(A, B)
_
V
T
F
V
T
F
_
= 0
_
R
11
R
12

1
0
0 R
22
0 0
_
_
_
_
_
V
T
1
F
V
T
2
F
V
T
1
F
V
T
2
F
_
_
_
_
= 0
Rank
_
(A, B)
_
= d = R
22
invertible = V
T
2
F = 0 = F span(V
1
)
= R
11
V
T
1
F +
1
V
T
1
F
= 0 = V
T
1
F
=
1
1
R
11
V
T
1
F
= F
= V
1
1
1
R
11
V
T
1
F
= LF
_
min norm solution
_
_
(AB
T
= BA
T
_
= L = L
T
18 / 42
th
Let P
v
= I V
1
V
T
1
and Q
v
= I P
v
be the orthogonal projectors
relative to node v then
A
v
F + B
v
F
= 0
_
P
v
F = 0
Q
v
F
+ L
v
F = 0
()
All self-adjoint realizations of 1 (the negative second derivative)
on with the vertex boundary conditions satisfy the following:
v 1 P
v
and Q
v
= I
d
v
P
v
(orthogonal projections) and
v 1 L
v
in Q
v
C
d
v
.
All f T(1)
H
2
(e) are described by (*)
at each nite vertex.
19 / 42
th
Quadratic form
The Quadratic form h of 1 is
h[f , f ] =
eE
_
e
df
ds
2
ds
vV
eE
(L
v
)
jk
f
j
(v)f
k
(v)
=
eE
_
e
df
ds
2
ds
vV
L
v
F, F,
where , is the standard Hermitian inner product in C
d
v
. The
domain of h consists of all f
eE
H
1
(e) such that P
v
F = 0.
20 / 42
th
Examples of b.c.
-type conditions
_
_
_
f (s) is continuous on
v
eE
v
df
ds
e
(v) =
v
f (v)
c
v
is the subset of the edges having v as a boundary point.
v
are real xed numbers
We describe the case for a node v of degree 3 (generalization is
easy)
21 / 42
th
Examples of b.c.
A
v
=
_
_
1 1 0
0 1 1
v
0 0
_
_
B
v
=
_
_
0 0 0
0 0 0
1 1 1
_
_
22 / 42
th
Examples of b.c.
A
v
=
_
_
1 1 0
0 1 1
v
0 0
_
_
B
v
=
_
_
0 0 0
0 0 0
1 1 1
_
_
A
v
B
T
v
=
_
_
0 0 0
0 0 0
0 0
v
_
_
The self-adjoint condition is satised i R
22 / 42
th
Examples of b.c.
A
v
=
_
_
1 1 0
0 1 1
v
0 0
_
_
B
v
=
_
_
0 0 0
0 0 0
1 1 1
_
_
A
v
B
T
v
=
_
_
0 0 0
0 0 0
0 0
v
_
_
L
v
=

v
d
v
_
_
1 1 1
1 1 1
1 1 1
_
_
22 / 42
th
Examples of b.c.
The Hamiltonian of the problem has the following h
h[f , f ] =
eE
_
e
df
ds
2
ds
vV
L
v
F, F
=
eE
_
e
df
ds
2
ds +
vV
v
[f (v)[
2
.
23 / 42
th
Examples of b.c.
h[f , f ] =
eE
_
e
df
ds
2
ds
vV
L
v
F, F
=
eE
_
e
df
ds
2
ds +
vV
v
[f (v)[
2
.
The case
v
0 corresponds to the Neumann-Kirchho
conditions
23 / 42
th
Examples of b.c.
h[f , f ] =
eE
_
e
df
ds
2
ds
vV
L
v
F, F
=
eE
_
e
df
ds
2
ds +
vV
v
[f (v)[
2
.
The case
v
0 corresponds to the Neumann-Kirchho
conditions
_
_
_
f (s) is continuous on
v
eE
v
df
ds
e
(v) = 0
h[f , f ] =
eE
_
e
df
ds
2
ds.
23 / 42
th
Examples of b.c.
-type conditions
_
_
v
The value of the derivative
df
e
ds
e
(s) is the same e c
v
eE
v
f
e
(v) =
df
ds
e
(v)
c
v
is the subset of the edges having v as a boundary point.
v
are real xed numbers
We describe the case for a node v of degree 3 (generalization is
easy)
24 / 42
th
Examples of b.c.
B
v
=
_
_
1 1 0
0 1 1
v
0 0
_
_
A
v
=
_
_
0 0 0
0 0 0
1 1 1
_
_
25 / 42
th
Examples of b.c.
B
v
=
_
_
1 1 0
0 1 1
v
0 0
_
_
A
v
=
_
_
0 0 0
0 0 0
1 1 1
_
_
A
v
B
T
v
=
_
_
0 0 0
0 0 0
0 0
v
_
_
25 / 42
th
Examples of b.c.
B
v
=
_
_
1 1 0
0 1 1
v
0 0
_
_
A
v
=
_
_
0 0 0
0 0 0
1 1 1
_
_
A
v
B
T
v
=
_
_
0 0 0
0 0 0
0 0
v
_
_
If
v
= 0 for some v then L
v
= 0.
L
v
= 0 v = h[f , f ] =
eE
_
e
df
ds
2
ds.
25 / 42
th
Examples of b.c.
B
v
=
_
_
1 1 0
0 1 1
v
0 0
_
_
A
v
=
_
_
0 0 0
0 0 0
1 1 1
_
_
A
v
B
T
v
=
_
_
0 0 0
0 0 0
0 0
v
_
_
If
v
,= 0 then B
v
is invertible and P
v
= 0 and Q
v
= I .
(L
v
)
i ,j
=
1
d
v
i , j
25 / 42
th
Examples of b.c.
h[f , f ] =
eE
_
e
df
ds
2
ds +
{vV|
v
=0}
1
eE
v
f (v)
2
.
The domain consists of all f (s)
e
H
1
(e) that have at each
vertex where
v
= 0 the sum of the vertex values along all the
incident edges is equal to 0.
26 / 42
th
Examples of b.c.
Dirichlet and Neumann conditions.
Dirichlet vertex conditions require that at each vertex the
boundary conditions impose f (v) = 0
The operator is decoupled in the sum of the negative second
derivative and
h[f , f ] =
eE
_
e
df
ds
2
ds
f H
1
().
The spectrum (1)
(1) =
_
n
2
2
l
2
2
[e (E), n Z 0
_
27 / 42
th
Examples of b.c.
Dirichlet and Neumann conditions.
Under Neumann vertex conditions no restriction on the value of
the function at vertices are required. The derivative at the vertices
are instead required to be zero. The operator is decoupled in the
sum of the negative second derivative and
h[f , f ] =
eE
_
e
df
ds
2
ds
f H
1
(), as for the Dirichlet case, but on a larger domain
27 / 42
th
Symmetric vertex conditions: a classication
We want to classify the cases for which the conditions
_
PF = 0
QF
+ LF = 0
()
are invariant under the action of the symmetric group of the
coordinate permutations. ( again we drop the subscript v)
28 / 42
th
We want to classify the cases for which the conditions
_
PF = 0
QF
+ LF = 0
()
are invariant under the action of the symmetric group of the
coordinate permutations. ( again we drop the subscript v) The
invariant space of the permutation group is the one dimensional
space generated by the vector of entries equal to one
=
1
d
(1, , 1)
T
C
d
.
Then (*) are invariant under the action of this group i P, Q, and
L are. We have 4 possible cases.
28 / 42
th
1. P = 0, Q = I , L =
T
+I
= 0
-conditions
= = 0 Neumann-conditions
2. P = I , Q = 0 (L irrelevant) Dirichlet-conditions
3. P = I

T
, Q =

T
, L = Q -conditions
4. P =

T
, Q = I

T
, L = P ????
29 / 42
th
1. P = 0, Q = I , L =
T
+I
= 0
-conditions
3. P = I

T
, Q =

T
, L = Q -conditions
4. P =

T
, Q = I

T
, L = P ????
29 / 42
th
1. P = 0, Q = I , L =
T
+I
= 0
-conditions
3. P = I

T
, Q =

T
, L = Q -conditions
4. P =

T
, Q = I

T
, L = P ????
29 / 42
th
1. P = 0, Q = I , L =
T
+I
= 0
-conditions
3. P = I

T
, Q =

T
, L = Q -conditions
4. P =

T
, Q = I

T
, L = P ????
29 / 42
th
Are these all the possible cases?
30 / 42
th
Are these all the possible cases?
NO!!
30 / 42
th
Modelling (Neumann Schr odinger example)
d
fat graph (l
e
length of edge e)
Let
d
denote the fat graph and
= d p(s) where p(s) > 0 is a
function of the arc length that can
be discontinuous at the vertices.
Each vertex neighbouring is
contained in a ball of radius d and
star-shaped with respect to a smaller
ball of diameter d.
31 / 42
th
On
d
we dene the Schrodinger operator
H
d
(A, q) =
_
1
i
A(s)
_
2
+ q(s)
with Neumann conditions on
d
(q scalar electric and A vector
magnetic potentials
32 / 42
th
On
d
we dene the Schrodinger operator
H
d
(A, q) =
_
1
i
A(s)
_
2
+ q(s)
with Neumann conditions on
d
(q scalar electric and A vector
magnetic potentials
H(A, q)f (s
e
) =
1
p
_
d
ds
e
iA
t
e
(s)
_
p
_
d
ds
e
iA
t
e
(s)
_
f + q
e
(s)f
where A
t
e
is the tangential component of A and q
e
is the
restriction of q to the graph.
32 / 42
th
Boundary conditions at the vertices
f is continuous through each vertex
{k|ve
k
}
p
k
_
df
k
ds
k
iA
t
k
f
k
_
(v) = 0
p
k
function that gives the width of the tube around e
k
. The values
of p
k
(v) at the same vertex can be dierent for dierent e
k
33 / 42
th
Boundary conditions at the vertices
f is continuous through each vertex
{k|ve
k
}
p
k
_
df
k
ds
k
iA
t
k
f
k
_
(v) = 0
p
k
function that gives the width of the tube around e
k
. The values
of p
k
(v) at the same vertex can be dierent for dierent e
k
Theorem For n = 1, 2, . . .
lim
0
n
(H
d
(A, q)) =
n
(H(A, q)) ,
where
n
is the n-th eigenvalue counted in increasing order
(accounting multiplicity)
33 / 42
th
Summary
We have dened the analytical structure of a Quantum Graph
the metric properties
the operators
the boundary conditions
We will focus in the last part on nite graphs and we will give
some hints on:
Solution of dierential equations
Eigenvalues problems
34 / 42
th
Summary
the operators
some hints on:
34 / 42
th
Summary
the operators
some hints on:
34 / 42
th
Summary
the operators
some hints on:
34 / 42
th
Summary
the operators
some hints on:
We assume that at the vertices of degree d = 1 we have Dirichlet
conditions.
34 / 42
th
Summary
the operators
some hints on:
We assume that at the vertices of degree d = 1 we have Dirichlet
conditions. Only Neumann-Kirchho conditions for sake of
simplicity
34 / 42
th
Dierential equations
Given a function g(s) L
2
() , we want to compute the solution
of the problem
min
u
h[u, u] g, u
with u(s) H
1
().
On each edge e c we solve the local problem and we denote

by u
e
(s
e
) the solution.
Using the Neumann-Kirchho conditions and the values of the

derivative at the vertices, we form and solve an algebraic
system on the vertices.
35 / 42
th
2
of the problem
min
u
h[u, u] g, u
with u(s) H
1
().

by u
e
(s
e
) the solution.

35 / 42
th
2
of the problem
min
u
h[u, u] g, u
with u(s) H
1
().

by u
e
(s
e
) the solution.

35 / 42
th
2
of the problem
min
u
h[u, u] g, u
with u(s) H
1
().

by u
e
(s
e
) the solution.

We have described an abstract Domain Decomposition approach.
35 / 42
th
Finite dimensional problem
Let assume of having a nite-element approximation. We subdivide
each edge forming a chain made of node of degree 2 and we build
the usual hat functions extending them to the vertices.
36 / 42
th
Finite dimensional problem
Let assume of having a nite-element approximation. We subdivide
each edge forming a chain made of node of degree 2 and we build
the usual hat functions extending them to the vertices.
36 / 42
th
The matrix
Reordering the nodes such that the internal nodes in a edge are
consecutive and the vertices are at the end we have a matrix
M =
_
M
11
M
12
M
T
12
M
22
_
where M
11
is block diagonal M
22
is diagonal and M
12
holds the
links between the edges.
37 / 42
th
Example
Graph (e edge and v vertex)
_
_
M
1
c
1
M
2
c
2
M
3
c
3
M
4
c
4
c
T
1
c
T
2
c
T
3
c
T
4
b
_
_
38 / 42
th
Eigenvalue Problem
The analysis of the spectrum of the self-adjoint operators is more
subtle. Innite quantum graphs can have Hamiltonian with
continuous part of the spectrum. However, for nite quantum
graphs we can have a better situation:
39 / 42
th
Eigenvalue Problem
The analysis of the spectrum of the self-adjoint operators is more
subtle. Innite quantum graphs can have Hamiltonian with
continuous part of the spectrum. However, for nite quantum
graphs we can have a better situation:
Theorem
Let a nite quantum graph with nite edges equipped with an
Hamiltonian given by negative second derivative along the edges
and vertex conditions
()
_
P
v
F = 0
Q
v
F
+ L
v
F = 0
Then the spectrum (1) is discrete.
39 / 42
th
An interesting connection
1f = f f L
2
()
Let e be an edge identied by the two vertices v and w of length
l
e
. If ,= n
2
2
l
2
e
with n Z 0 then
f
e
=
1
sin
l
e
_
f
e
(v) sin
(l
e
s) + f
e
(w) sin
s
_
40 / 42
th
1f = f f L
2
()
l
e
. If ,= n
2
2
l
2
e
with n Z 0 then
f
e
=
1
sin
l
e
_
f
e
(v) sin
(l
e
s) + f
e
(w) sin
s
_
Substituting
f

e
(v) =
l
e
sin
_
f
e
(w) f
e
(v) cos l
e
_
in (*) and eliminating the derivatives we compute a system of
algebraic relations
40 / 42
th
1f = f f L
2
()
l
e
. If ,= n
2
2
l
2
e
with n Z 0 then
f
e
=
1
sin
l
e
_
f
e
(v) sin
(l
e
s) + f
e
(w) sin
s
_
Substituting
f

e
(v) =
l
e
sin
_
f
e
(w) f
e
(v) cos l
e
_
in (*) and eliminating the derivatives we compute a system of
algebraic relations
T()F = 0
40 / 42
th
Theorem
,= n
2
2
l
2
e
with n Z 0 belongs to the spectrum of 1 i
zero belongs to the spectrum of T()
41 / 42
th
Theorem
,= n
2
2
l
2
e
This results connects quantum graph theory to combinatorial
graph theory
41 / 42
th
Theorem
,= n
2
2
l
2
e
What about = n
2
2
l
2
e
?
41 / 42
th
Theorem
,= n
2
2
l
2
e
What about = n
2
2
l
2
e
?
If we have only Dirichlet conditions they are the only eigenvalues,
otherwise ??????
41 / 42
th
Conclusion
Quantum graphs are independent from the embedding space

and his make them a good candidate to model complex
phenomena depending on many variables.
The process of model reduction can be tricky
not always self-adjoint operators
continuous spectrum issues
ray edges ....
Non linear operators: p-Laplacian,..
Integro-dierential operators (fractional derivatives and

visco-elasticity, ....)
Interesting numerical linear algebra and potential high

parallelism
Strong interaction between dierent expertises ...

42 / 42
th
Conclusion

ray edges ....


parallelism

42 / 42
th
Conclusion

ray edges ....


parallelism

42 / 42
th
Conclusion

ray edges ....


parallelism

42 / 42
th
Conclusion

ray edges ....


parallelism

42 / 42
th
Conclusion

ray edges ....


parallelism

42 / 42

Quantum Graphs PDF

Cargado por

Información del documento

Descripción original:

Título original

Derechos de autor

Formatos disponibles

Compartir este documento

Compartir o incrustar documentos

Opciones para compartir

¿Le pareció útil este documento?

¿Este contenido es inapropiado?

Copyright:

Formatos disponibles

Quantum Graphs PDF

Cargado por

Copyright:

Formatos disponibles

An introduction to Quantum Graphs theory

Motivations and Modelling

Graphs, metric graphs, and quantum graphs

Self-adjoint Hamiltonians and boundary conditions

Waves and eigenvalues problems

Numerical issues and domain decomposition

Open problems i.e.

Motivations and Modelling

Graphs, metric graphs, and quantum graphs

Self-adjoint Hamiltonians and boundary conditions

Waves and eigenvalues problems

Numerical issues and domain decomposition

Open problems i.e. The things I have not understood yet!!!

Dicult to have a decent

Irregular solution (corners)

Given an Hamiltonian on the fat

Dicult to have a decent

Irregular solution (corners)

Given an Hamiltonian on the fat

Dicult to have a decent

Irregular solution (corners)

Given an Hamiltonian on the fat

Dicult to have a decent

Irregular solution (corners)

Given an Hamiltonian on the fat

Dicult to have a decent

Irregular solution (corners)

Given an Hamiltonian on the fat

A Combinatorial Graph is dened by a set 1 = v

A graph is a Metric Graph if at each edge e is assigned a

A Combinatorial Graph is dened by a set 1 = v

A graph is a Metric Graph if at each edge e is assigned a

A Combinatorial Graph is dened by a set 1 = v

A graph is a Metric Graph if at each edge e is assigned a

We can remove vertices of degree 2 after we fuse the 2 edges

The graph is a topological manifold (or a 1D simplicial

is provided with a global metric and the distance between

is NOT necessarily embedded in a Euclidean space (R

We can remove vertices of degree 2 after we fuse the 2 edges

The graph is a topological manifold (or a 1D simplicial

is provided with a global metric and the distance between

is NOT necessarily embedded in a Euclidean space (R

We can remove vertices of degree 2 after we fuse the 2 edges

The graph is a topological manifold (or a 1D simplicial

is provided with a global metric and the distance between

is NOT necessarily embedded in a Euclidean space (R

We can remove vertices of degree 2 after we fuse the 2 edges

The graph is a topological manifold (or a 1D simplicial

is provided with a global metric and the distance between

is NOT necessarily embedded in a Euclidean space (R

f is continuous through each vertex

f is continuous through each vertex

We have dened the analytical structure of a Quantum Graph

the metric properties

the boundary conditions

Solution of dierential equations

We have dened the analytical structure of a Quantum Graph

the metric properties

the boundary conditions

Solution of dierential equations

We have dened the analytical structure of a Quantum Graph

the metric properties