Poisson Processes and Markov Chains in Mathematics: Homework 4, Assignments of Cryptography and System Security

Solutions to homework problems 3.3, 3.5, and 3.7 in a mathematics 171 course. Topics covered include poisson random variables, independence of random variables, irreducible markov chains, and mean transit times.

Typology: Assignments

Pre 2010

Uploaded on 09/17/2009

koofers-user-ds2
koofers-user-ds2 🇺🇸

10 documents

1 / 2

Toggle sidebar

This page cannot be seen from the preview

Don't miss anything!

bg1
Math 171 Homework #4
3.3
a. Since Xand Yare Poisson, then XtiXsiand YtiYsiare independent
random variables in non-overlapping intervals tisi. Thus XtiXsi+YtiYsi
are independent.
If events in Xand Yoccur one at a time, then X+Yoccurs one at a time.
The only other possibility is that an event in Xand Yoccur at the same
time. But this occurs with probability limt0λ1λ2(∆t)2.
Events in Zoccur with rate (λ1+λ2).
b. We get λ1
λ1+λ2.
c. The probability that someone has called phone 1 is 1 eλ1t. The prob-
ability that someone has called both phones is then (1 eλ1t)(1 eλ2t) =
{Tt}.
3.5
A=
1 2
11 1
244
=1 1
140 0
054/5 1/5
1/51/5
Pt=etA
=1 1
140 0
0e5t4/5 1/5
1/51/5
=4/5 1/5
4/5 1/5+e5t1/51/5
4/5 4/5
3.7
Consider two arbitrary states i,j. Because the system is irreducible, we know
that there exit z1, z2, . . . , xkSwith α(x, z1), α(z1, z2), . . . , α(zk1, zk), α(zk, y)
all strictly positive. We can divide time into (k+ 1) subintervals and write
P{Xt=j|X0=i} P{Xt/(k+1) =z1|X0=i}P{X2t/(k+1) =z2|Xt/(k+1) =z1} · · ·
· · · P{Xkt/(k+1) =zk|X(k1)t/(k+1) =zk1}P{Xt=j|Xkt/(k+1) =zk}
α(i, z1)(t/(k+ 1))α(z1, z2)(t/(k+ 1)) · · · α(zk, j )(t/(k+ 1))
0
3.8
pf2

Partial preview of the text

Download Poisson Processes and Markov Chains in Mathematics: Homework 4 and more Assignments Cryptography and System Security in PDF only on Docsity!

Math 171 Homework #

a. Since X and Y are Poisson, then Xti − Xsi and Yti − Ysi are independent random variables in non-overlapping intervals ti −si. Thus Xti −Xsi +Yti −Ysi are independent. If events in X and Y occur one at a time, then X + Y occurs one at a time. The only other possibility is that an event in X and Y occur at the same time. But this occurs with probability lim∆t→ 0 λ 1 λ 2 (∆t)^2. Events in Z occur with rate (λ 1 + λ 2 ).

b. We get (^) λ 1 λ+^1 λ 2.

c. The probability that someone has called phone 1 is 1 − e−λ^1 t. The prob- ability that someone has called both phones is then (1 − e−λ^1 t)(1 − e−λ^2 t) = {T ≤ t}.

A =

[

1 2 1 − 1 1 2 4 − 4

]

[

] [

] [

]

Pt = etA

=

[

] [

0 e−^5 t

] [

]

[

]

  • e−^5 t

[

]

Consider two arbitrary states i,j. Because the system is irreducible, we know that there exit z 1 , z 2 ,... , xk ∈ S with α(x, z 1 ), α(z 1 , z 2 ),... , α(zk− 1 , zk), α(zk, y) all strictly positive. We can divide time into (k + 1) subintervals and write

P{Xt = j|X 0 = i} ≥ P{Xt/(k+1) = z 1 |X 0 = i}P{X 2 t/(k+1) = z 2 |Xt/(k+1) = z 1 } · · ·

· · · P{Xkt/(k+1) = zk|X(k−1)t/(k+1) = zk− 1 }P{Xt = j|Xkt/(k+1) = zk} ≥ α(i, z 1 )(t/(k + 1))α(z 1 , z 2 )(t/(k + 1)) · · · α(zk, j)(t/(k + 1)) ≥ 0

a. The equilibrium distributin is the left eigenvector of A corresponding to eigenvalue 0. This is

[

]

b. The rate at which the system leaves state 1 is α(1) = 3, so the expected time to leave that state is 1/3.

c. Remove state four from the matrix to get

A˜ =

1 2 3 1 − 3 1 1 2 0 − 3 2 3 1 2 − 4

Then [− A˜]−^1 ¯1 = ¯1, so the mean transit time is 1.