Introduction to Graph Theory Dr. James Burk Course: Math 400: {Graph Theory} Office: Science Building 106 Email: [email protected]Fall 2016 Dr. James Burk Introduction to Graph Theory A problem to solve A map of Searcy .... Dr. James Burk Introduction to Graph Theory A problem to solve A UPS driver has ten stops to make.... Dr. James Burk Introduction to Graph Theory A problem to solve 15 possible roads to take between stops.... Dr. James Burk Introduction to Graph Theory Notes Notes Notes Notes
49
Embed
Introduction to Graph Theory - Matrix Analysis · Graph Theory - An Introduction Definition Agraphis a collection of points calledverticesand lines called edgesthat connect someof
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
During a walk through the city, a question was asked:
Is it possible to leave home, cross each bridge exactly once, and thenreturn home?
Dr. James Burk Introduction to Graph Theory
Graph Theory - An Introduction
The “bridge problem” can be rephrased as:
Is it possible to start at one vertex, and “walk” along the edges sothat:
1 all edges are crossed exactly once, and2 the “walk” ends at the same vertex it started.
If this is possible, the graph is called an Eulerian graph, afterLeonard Euler (pronounced “Oiler”)
Dr. James Burk Introduction to Graph Theory
Exercises – page 1 of 2
1 For each graph below, find the number of edges, the number ofvertices, the degree of each vertex. Verify that the “Sum ofDegrees Theorem” is true for each graph.
What if, given a graph, we are interested in traversing all edges, butwe don’t care if the starting and ending points are the same?
DefinitionAn Euler walk is a path that uses each edge exactly once, but endsat a point different from its starting point.
Example:
Dr. James Burk Eulerian Graphs
Eulerian Walks
Theorem (Euler Walk Theorem)
A connected graph contains an Euler Walk when all but two verticeshave even degree.
– that is, two with odd degree, and the rest with even degree.
The Euler walk will begin and end at the vertices with odd degree.
Dr. James Burk Eulerian Graphs
Eulerian Walks
Example: A campground has 7 campsites, with 11 bike trails.
Is it possible to start at one campground, bike each trail once, andend at the same campground? NOIs it possible to start at one campground, bike each trail once, andend at another campground? YES
Dr. James Burk Eulerian Graphs
Homework – page 1 of 2
1 Is it possible to leave one city, travel all the roads exactly once,and return to the starting city? Why or why not?
Dover
Evanston
Caldwell
Fairmont
Burley
Alameda
Grange
2 Does the graph below contain an Euler walk? Why or why not? Ifit does, find one.
How many edges are in the complete graph with n vertices?
4 vertices: # of edges: 6 degree of each vertex: 35 vertices: # of edges: 10 degree of each vertex: 46 vertices: # of edges: 15 degree of each vertex: 57 vertices: # of edges: 21 degree of each vertex: 6
Dr. James Burk Hamiltonian Graphs
Hamiltonian Graphs
Why talk about complete graphs right now?
Because there are always more than enough edges to satisfy theHamilton Graph theorem:
Theorem (Hamilton Circuit Theorem)
Suppose we have a connected graph that has no multiple edges, andhas n vertices. If every vertex has degree at least n
2 , then the graph isHamiltonian.
So every complete graph is definitely Hamiltonian!!WARNING: This theorem only says a Hamiltonian circuit exists. Itdoesn’t tell us how to find it. So, how can we find them?
Dr. James Burk Hamiltonian Graphs
Finding Efficient Routes
We’ll actually do something much better.
In many contexts, the edges of a graph are defined in such a way thatsome edges are better (or worse) than others.
For example, a UPS driver might prefer to drive a shorter road than alonger road between cities.
We could keep track of these differences with a labeled graph.
ABSTRACTIn [16] a methodology based on Game Theory for the anal-ysis of gene expression data is studied. Roughly speaking,the starting point is the observation of a ‘picture’ of gene ex-pressions in a sample of cells under a biological condition ofinterest, for example a tumor. Then, Game Theory plays aprimary role to quantitatively evaluate the relevance of eachgene in regulating or provoking the condition of interest, tak-ing into account the observed relationships in all subgroupsof genes. In this paper, an alternative model based on min-imum cost spanning tree representation of gene expressiondata has been introduced. One of the main characteristicsof this model is the possibility to avoid the dichotomizationtechnique required for microarray games introduced in [17].
1. INTRODUCTIONGene expression occurs when genetic information con-
tained within DNA is transcripted into messenger ribonu-cleic acid (mRNA) molecules and then translated into theproteins. So, estimates of the amount of mRNA for a givengene yields a quantitative evaluation of the amount of pro-teins coded by that gene, that is its gene expression level.Nowadays, DNA microarray technology is available for tak-ing ‘pictures’ of gene expressions. Within a single experi-ment of this sophisticated technology, the ratio of expressionlevels of thousands of genes under two di!erent experimentalconditions can be estimated. The expression level of eachgene collected from a cell under a biological condition ofinterest (e.g. a tumor), can be compared to the expressionlevel of the same gene collected from a control cell (usuallyan healthy one), and the ratio of expression levels between
Permission to make digital or hard copies of all or part of this work forpersonal or classroom use is granted without fee provided that copies arenot made or distributed for pro£t or commercial advantage and that copiesbear this notice and the full citation on the £rst page. To copy otherwise, torepublish, to post on servers or to redistribute to lists, requires prior speci£cpermission and/or a fee.GameNets’06, October 14, 2006, Pisa, Italy.Copyright 2006 ACM 1-59593-507-X/06/10...$5.00
the two conditions can be computed for each gene. If k > 0experiments are performed, each gene i in the set of stud-ied genes N can be associated to a k-dimensional vector ofreal-valued ratio of expression levels, that we will refer asthe expression vector of gene i, for each i ! N .
In [17] the class of microarray games has been introduced.Via a dichotomization technique applied to gene expressiondata, it is constructed a game whose characteristic functiontakes values on the interval [0, 1]. The objective of sucha game is to stress the relevance (‘su"ciency’) of groups ofgenes in relation to a specific biological condition or responseof interest (e.g. a disease of interest). In that paper, it hasbeen discussed the possibility of applying game-theoreticaltools that can take into account the relationships which existamong genes, like the Shapley value. The highest Shapleyvalues of the game should point to the most influential genes,so that it could be useful as a hint for pointing at the genesthat mostly deserve further investigation.
Microarray games may be useful to describe the magni-tude of the association of a given biological condition of theoriginal cell (e.g. a tumor) with a given gene expression cate-gory (e.g., ‘abnormal expression’, ‘down regulation’, ‘strongvariability’ etc.) of genes realized in each coalition of genes,i.e. each subset of N . As we already mentioned, in order toset up a microarray game, the real-valued ratio of expressionlevels must firstly be turned into expression categories usingappropriate cuto!s on the continuous domain.
A key issue in the definition of microarray games is the op-erational definition of association between a gene expressioncategory and a biological condition realized in a nonemptycoalition S " N . Given a set of genes N collected from acell, the authors in [17] claimed that a su"cient conditionto realize in a nonempty coalition S " N the associationbetween the expression category and a biological conditionof the cell is that all the genes which present the expres-sion category belong to coalition S (su!ciency principle forgroups of genes). For example, if the expression categoryis ‘over expression’ and the biological condition of the cellis ‘tumor’, then the association between the tumor and theover expression category is realized in a coalition S " N ,S #= $, if all the over expressed genes in the tumor cell be-long to S. Subsequently, given a set of observed cells, themagnitude of the association is measured, in each coalitionS, as the observed frequency of associations realized in Saccording to the su"ciency principle. Following the exam-ple above, if a nonempty coalition of genes S " N contains
Dr. James Burk Trees
Finding trees inside a graph
So, how do we find a spanning tree inside a given graph?
Just like with Hamilton circuits, we’ll do something better:
We’ll find the best spanning tree in a graph with labeled edges.
This will be called the minimal spanning tree for the graph.
1 How many edges are in a tree that has 15 vertices?2 How many vertices are in a tree that has 15 edges?3 Find all spanning trees in the following graph:
4 Find all spanning trees in the following graph:
Dr. James Burk Trees
Homework - page 2 of 2
4 Draw two different spanning trees from the following graph:
5 Use Kruskal’s Algorithm to find the minimal spanning tree insidethe following labeled graph. What is the “total cost”?
15
11
14
10 31
30
33 18
27
20
1223
Dr. James Burk Trees
Homework - page 2 of 2
6 Use Kruskal’s Algorithm to find the minimal spanning tree in thefollowing labeled graph. What is the “total cost”?
In this table, an X means that there is one student in both clubs...
Skiing SGA Debate NHS Paper Service Dem. Rep.Skiing – X XSGA X – X X XDebate X – X X XNHS X X – X X XPaper X X – X XService X X X – X XDem. X X X –Rep. X X X –
Dr. James Burk More Map Coloring
Scheduling
Let’s organize this in a graph:
Dr. James Burk More Map Coloring
Scheduling
Let’s organize this in a graph:
Dr. James Burk More Map Coloring
Scheduling
Let’s organize this in a graph:
An edge between vertices means that these clubs can’t meet at thesame time.
Will 2 meeting times be enough? Is the graph 2-colorable? NO!
Dr. James Burk More Map Coloring
Scheduling
Will 3 meeting times be enough? Is the graph 3-colorable?
I don’t think so!
Dr. James Burk More Map Coloring
Scheduling
Will 4 meeting times be enough? Is the graph 4-colorable?
YES!
Dr. James Burk More Map Coloring
Another application – traffic management
Consider an intersection:
There are 6 lines of traffic: A, B, C, D, E, FSome of the traffic lines can’t move without risk of collision.For example, A and D.How many different cycles of traffic lights will be needed?
Eight political committees must meet on the same day, but some membersare on more than one committee. So those committees that have members incommon cannot meet at the same time. An “X” in the table below indicatesthat the two corresponding committees share a common member. Usegraph coloring to determine the least number of meeting times that will benecessary so that all members can attend their committee meetings.
A B F J Ed. He FA HoApprop. – X X XBudget – X XFinance X – X X XJudiciary X X – X XEd. – X XHealth X X X –For. Aff. X X X –Housing X X X –