Slides
Hour 1
The
first hour will cover a general overview of graph coloring. Topics
that will be discussed include a series of applications that fit
the domain of graph coloring (such as trying to color a planar
map and scheduling problems), definitions and terminology that
is associated with graph coloring (such as the chromatic number
and the k-coloring of a graph) and existing results of graph coloring.
Some examples of existing results which will be looked at include:
- A Complete graph
of n vertices, Kn, requires at least n colors.
- The chromatic
number of a cycle of length n, Cn, is 2 if the cycle is even.
- A complete Bipartite
graph with m sets of vertices and n sets of vertices can be
colored using at most two colors.
Examples of both edge
and vertex colored graphs will also be given.
Hour 2
During
the remainder of the first class there will be a quick introduction
to Bipartite and Planar Graphs. This will include an algorithm
for determining if a given graph is two colorable (this algorithm
was already seen in Algorithms since determining if a graph is
bipartite is necessary and sufficient for showing a graph to be
2-colorable). The notion of 4-colorability of planar graphs will
also be investigated. I will include a history of the problem
along with the stumbling blocks that have been encountered along
the way. The end of this discussion will include looking at the
use of computers in proving the result.
Hours 3 and 4
This
talk will introduce two problems that have been investigated in
graph coloring. The first problem that will be investigated will
be to determine if graph is 3-colorable. The second problem will
be to find a minimum coloring of a graph. Two different approaches
will be discussed to solving this problem. The first is a contraction
algorithm called the Recursive-Largest-First (RLF). The second
is a sequential coloring algorithm called the Largest First (LF)
algorithm. In each case pseudo-code, and example and timings will
be discussed. It is known that these problems are NP-complete.
For example, problem 2 can be shown to be NP-complete by reducing
the 3-satisfiability problem (3-SAT) to Colorability. I would
also like to find a good algorithm for edge colorings. When I
find such an algorithm that I find interesting I will be looking
to add it to the talk as well.
Time Permitting
If
there is time that remains a number of other areas can be discussed
including plane-coloring, the coloring of hypergraphs and non-deterministic
methods of approximation.