The Wooly Hat
The Woolly Hat
The Toque
The Tuque
The Waldo Hat
An Update
This is joint work with
Leah Berman
Hiroki Koike
Alejandra Ramos-Rivera
Primož Potočnik
Primož Šparl
Elias Mochan
Mod N
A
B
C
a
b
c
d
This means:
Vertices are all Ai, Bi, Ci, for numbers i mod n
Edges are all:
{Ai, Ai+a}
{Ai, Bi}
{Ai, Ci}
{Bi, Ci+b}
{Bi, Ci+c}
{Bi, Ci+d}
WHn(a,b,c,d)
Classification of edge-�transitive d-valent k-circulants:
k=1
Generalized Petersen Graphs
k=2
1
K4, K3,3
a
1
1
K4
a
b
{3,6} torus maps
Classification of edge-�transitive 3-valent k-circulants:
Classification of edge-�transitive 3-valent k-circulants:
k=3
Pappus
8-cage
{3,6}3,3
K3,3
Kovàcs, Kutnar, Marušič, Wilson
N=any, a2 = ±1 mod N
N=2m, a = m±1
a
1
Classification of edge-�transitive 4-valent k-circulants:
k = 1
k = 2
a
1
r
Rose Window Graphs: Four families
Kovacs, Kuzman, Malnič, Wilson
a
b
c
Three families and three sporadics
Kovàcs, Kutnar, Marušič
Tricirculant tetravalent graphs
Mod N
CMO, Oaxaca 2014
Leah Berman
Hiroki Koike
Elias Mochan
MAGMA trials on WHn(a,b,c,d) for n up to n=80:
(1) For which values of the parameters is the graph edge-transitive?
(2) For which values of the parameters is the graph vertex-transitive?
(3) For which values of the parameters is the graph a LR structure?
LOTS. At least n/5 for every even value of n
None. Not one.
Many. At least half of the vertex-transitive ones
Questions about WHn(a,b,c,d) and answers from
First Family:
If n and c are even, while a, b, d are all odd, d-b = 2a and a+2c = b+2d, then WHn(a,b,c,d) is vertex-transitive.
If d-b = 2a, then a+2c = b+2d is equivalent to
2c = 3(a+b).
Theorem A
Theorem A explains 99% of all examples of V-T WH graphs
Which leaves exactly two sporadic cases:
Theorem B explains another .98%
If k is odd, then WH4k(2, k-2, 0, k+2) is vertex-transitive.
At SIGMAP 2018 (Morelia):
WH8(2, 1, 0, 5) and WH8(2, 1, 4, 5)
Theorem B
A
B
C
a
b
c
d
There is always a symmetry μ interchanging Ai with A-i and Bi with C-i.
0
-b+c
b
d-b
c-d
c
d
d-b+c
c-d+b
b-c
b-d
d-c
d+b-c
a
b
c
d
Suppose the graph is vertex-transitive but not edge-transitive
A
C
B
Each of the canonical hexes is
a
b
c
d
There must be a hex like that through A0B0.
A
C
B
That hex must be AAABCB
Tracing, we see that
d-b = 2a
Now consider this alternating 8-cycle:
Where there is one, there must be many.
A0
Aa
Ab
Aa+b
B0
Ba
Cb
Ca+b
A_: b-a, b+a, b+3a
C_: 2b-c, 2b-c+2a, 2b-c+4a
A_: c-b-3a, c-b-a, c-b+a
B_: 2c-2b-4a, 2c-2b-2a, 2c-2b
A B:
3(a+b) = 2c
C A:
3(a+b) = 2c
C B:
4(a+b) = 4c
. . but that they know so much that ain’t so.
Josh Billings
It ain’t so much that folks is iggerant
a
b
c
d
C
A
B