Here are slides from the presentation Kenny Ascher and I gave at Brown University. The work was part of research done by Caleb Holtzinger, Kenny Ascher, Professor Michael Falk and myself at Northern Arizona University, funded by the National Science Foundation.
It should be noted that an edge-joint partition of a graph is some partition of a graph's edges such that if an edge, e (between vertex a and b), is in a partition then if for any vertex c not equal to a or b, c is in the neighborhood of a AND b then ac and bc are also in the same partition as e. If a graph is simple this is the same as saying that if an edge is in a partition any K_3's that contain this edge are also in said partition.
A graph is edge-joint maximal if there exists no non-trivial edge-joint partition (i.e. the only edge-joint partition is of G with the empty set).
A maximal edge-joint component, A, of a graph, G, is an edge-joint maximal sub-graph, A, such that the partition {E(A), E(G-A)} is an edge-joint partition.
6. Topology of
Graphic
Hyperplane
Arrangements
Kenneth
Ascher &
Donald
Mathers
Introduction
Definitions
Example
More
Definitions
Graphs
Algebra
O.S. Algebra
Resonance
Varieties
Results
Resonance
Varieties
Polymatroid
Example
Future
Future
End
Orlik-Solomon Algebra
We are interested in the complement of A ,
M = C −
n
i=1
Hi (or in complex projective space)
For each Hi ∈ A we have a corresponding basis vector, ei .
Let E = ∧(e1,··· ,e ) be the exterior algebra
Define ∂ : E p → E p−1 by:
∂(ei1 ∧···∧eip ) =
p
∑
k=1
(−1)k−1
ei1 ∧···∧eik
∧···∧eip
Notation: S = (i1,··· ,ip), denote ei1 ∧···∧eip as eS
11. Topology of
Graphic
Hyperplane
Arrangements
Kenneth
Ascher &
Donald
Mathers
Introduction
Definitions
Example
More
Definitions
Graphs
Algebra
O.S. Algebra
Resonance
Varieties
Results
Resonance
Varieties
Polymatroid
Example
Future
Future
End
Resonance Varieties
We are interested in the dimension of the (first) resonance
variety and have the following theorem:
Theorem
Given a graphic arrangement, AΓ, let B be the arrangement
formed by removing all hyperplanes in A which are not
contained in a K3. Call this resulting graph associated to B, H.
Then, dim(R1(A )) = e(H)−c(H), where e(H) denotes the
number of edges in H, and c(H) denotes the number of
maximal edge-joint components.
Corollary
Given a 2-connected, parallel-indecomposable graph Γ, such
that each edge is contained in a K3, dim(R1(A )) = e(Γ)−1.
13. Topology of
Graphic
Hyperplane
Arrangements
Kenneth
Ascher &
Donald
Mathers
Introduction
Definitions
Example
More
Definitions
Graphs
Algebra
O.S. Algebra
Resonance
Varieties
Results
Resonance
Varieties
Polymatroid
Example
Future
Future
End
Polymatroids
We’ve already established that the resonance variety is a union
of linear subspaces.
Let’s assume R1(A ) = L1 ∪L2 ···∪Lk M1 ∪···∪Mn
Li ↔ K3
Mi ↔ K4
The polymatroid is a function that assigns to each
subspace of R1, the dimension of its span
Using our theorem on dimension of the span of the resonance
variety, we can calculate the polymatroid of a graph containing
no 4-cliques.
14. Topology of
Graphic
Hyperplane
Arrangements
Kenneth
Ascher &
Donald
Mathers
Introduction
Definitions
Example
More
Definitions
Graphs
Algebra
O.S. Algebra
Resonance
Varieties
Results
Resonance
Varieties
Polymatroid
Example
Future
Future
End
Polymatroid
Let Γ be a K4-free graph (a decomposable arrangement)
Each K3 contributes a local component, a linear space of
dimension 2, so we are interested in the dimension of the
span of the union of any components
We call the dimension of the span degenerate if it is “less
than expected“
Wheel graphs represent “minimal degenerate sets“
We can determine the polymatroid of these arrangements by
looking at subgraphs which are wheel graphs.