Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                
Scribd
Upload
83 views

Computer Network Assignment Help

The document provides solutions to problems related to modeling networks and processes as Markov chains. It solves problems involving computing centrality measures for different network structures, modeling disease spread and a fishing process as Markov chains, and identifying recurrent and transient states. The solutions include deriving transition probabilities and drawing Markov chains to represent the given processes.

Uploaded by

Computer Network Assignment Help
Copyright
© © All Rights Reserved
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
83 views

Computer Network Assignment Help

The document provides solutions to problems related to modeling networks and processes as Markov chains. It solves problems involving computing centrality measures for different network structures, modeling disease spread and a fishing process as Markov chains, and identifying recurrent and transient states. The solutions include deriving transition probabilities and drawing Markov chains to represent the given processes.

Uploaded by

Computer Network Assignment Help
Copyright
© © All Rights Reserved
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 11

For any Assignment related queries, Call us at : -  

+1 678 648 4277


You can mail us at : - support@computernetworkassignmenthelp.com or
reach us at : - www.computernetworkassignmenthelp.com/

Computer Network Assignment Help


Problem 2.1
Let the directed graph G be a ring: node i is connected to i + 1 if i < m and m is connected
to 1. Compute both eigenvalue centrality and Katz centrality (with = 1). Comment on your
result. Do the same for a k-regular undirected network (i.e., an undirected network in
which every vertex has degree k). You may find the steps outlined in Newman Problem 7.1
helpful. Comment on your result.

Solution:
(a) Eigenvector Centrality: For an adjacency matrix A, the eigenvector with the highest
eigenvalue represents the eigenvector centrality of each node in A. G is an orthogo-nal
matrix with the highest eigenvalue 1 and corresponding eigenvector

Here we prove that G can not have an

Computer Network Assignment Help


Here we show that 1 is an eigenvalue of A because its an orthogonal matrix with unit
vectors

This means that all eigenvalues of G  1. But since G is orthogonal, all of its eigenvalues
lie on the complex plane wit odulus 1 which means

Naturally, all nodes have equal centrality (both eigenvector and katz) due to symmetry.

Computer Network Assignment Help


Setting x1 = ... = xn = x and c = 1 and as the attenuating factor, we can write

Thus, the centralities increase as k increases. All nodes have equal centrality (both
eigenvector and katz) due to symmetry.

Problem 2.2
[Problem 7.2 from Newman] Suppose a directed network takes the form of a tree with all
edges pointing inward towards a central vertex: (see figure in Newman). What is the
PageRank centrality of the central vertex in terms of the single parameter ↵ appearing in
the definition of PageRank and the geodesic distances di from each vertex i to the central
vertex?

Computer Network Assignment Help


Solution:
We want to computer the Pagerank centrality of xc of the central vertex and we assume
that C = 1. Notice that 3 nodes have distance 3 from c, 4 nodes have distance 2 from c
and 2 nodes have distance 1 from c. Generalizing from the pattern any node that is
distance d from c will have its centrality scaled with ↵ d number of times, towards the
computation of xc.

Therefore,

Computer Network Assignment Help


Problem 2.3

(a) Give an example of such a graph with 5 nodes.


(b) Compute the eigenvalue centrality. Explain your answer.
(c) Compute the Katz centrality with = 1. Explain your answer.

Solution:
(a) Any directed acyclic graph would do. Here is an example given below

Computer Network Assignment Help


(b) Since A is nilpotent, the characteristic polynomial is t n = 0 and all the eigenvalues i are
zero. So it does not make sense to look at the eigenvector centrality, defined as the
eigenvector associated with the largest eigenvalue, max.

(c) The katz centrality is computed iteratively using

. Picking ↵ = 0.3, for instance, and = 1 we get (after 4 iterations)

Computer Network Assignment Help


ckatz = [1 1.3 1.39 1.417 1.4251]

which provides a sensible ranking if we take incoming edges to indicate importance.

Problem 2.4

As flu season is upon us, we wish to have a Markov chain that models the spread of a flu
virus. Assume a population of n individuals. At the beginning of each day, each individual is
either infected or susceptible (capable of contracting the flu). Suppose that each pair (i, j), i
6= j, independently comes into contact with one another during the daytime with probability
p. Whenever an infected individual comes into contact with a susceptible individual, he/she
infects him/her. In addition, assume that overnight, any individual who has been infected for
at least 24 hours will recover with probability 0 <q< 1 and return to being susceptible,
independently of everything else (i.e., assume that a newly infected individual will spend at
least one restless night battling the flu)

(a) Suppose that there are m infected individuals at daybreak. What is the distribution
of the number of new infections by day end?
(b) Draw a Markov chain with as few states as possible to model the spread of the flu for
n = 2. In epidemiology, this is called an SIS (SusceptibleInfected-Susceptible) model.
(c) Identify all recurrent states.

Due to the nature of the flu virus, individuals almost always develop immunity after
contracting the virus. Consequently, we improve our model and assume that individuals
become infected at most one time. Thus, we consider individuals as either infected,
susceptible, or recovered.
(d) Draw a Markov chain to model the spread of the flu for n = 2. In epidemiology, this is
called an SIR (Susceptible-Infected-Recovered) model.
(e) Identify all recurrent states.

Solution:

Computer Network Assignment Help


(a) If m out of n individuals are infected, then there must be n m susceptible individuals. Each one
of these individuals will be independently infected over the course of the day with probability p =
1 - (1 - p )m.

Thus the number of new infections, I, will be a binomial random variable with parameters n m
and p. That is,

(b) Let the state of the SIS model be the number of infected individuals. For n = 2, the
corresponding Markov chain is illustrated below.

Computer Network Assignment Help


(c) The only recurrent state is the one with 0 infected individuals

(d) Let the state of the SIR model be (S,I), where S is the number of susceptible
individuals and I is the number of infected individuals. For n = 2, the corresponding
Markov chain is illustrated below

Computer Network Assignment Help


If one did not wish to keep track of the breakdown of susceptible and recovered
individuals when no one was infected, the three states free of infections could be
consolidated into a single state as illustrated below
(e) Any state where the number of infected individuals equals 0 is a recurrent state.
For n = 2,there are either one or three recurrent states, depending on the Markov
chain drawn in part(d).

Computer Network Assignment Help


Problem 2.5

There are n fish in a lake, some of which are green and the rest blue. Each day, Helen
catches 1 fish. She is equally likely to catch any one of the n fish in the lake. She throws back
all the fish, but paints each green fish blue before throwing it back in. Let Gi denote the
event that there are i green fish left in the lake.

(a) Show how to model this fishing exercise as a Markov chain, where Gi are the
states. Explain why your model satisfies the Markov property.
(b) Find the transition probabilities pij
(c) List the transient and recurrent states

Solution:

(a) The number of remaining green fish at time n completely determines all the relevant
information of the systems entire history (relevant to predicting the future state). Therefore
it is immediate that the number of green fish is the state of the system and the process has
the Markov property:

(b) For j > i clearly pij = 0, since a blue fish will never be painted green. For 0  i, j  k, we
have the following

(c) The state 0 is an absorbing state since there is a positive probability that the system
will enter it, and once it does, it will remain there forever. Therefore the state with 0
green fish is the only recurrent state, and all other states are then transient.

Computer Network Assignment Help

You might also like