Scribd
Upload
70 views15 pages

Stochastic Models: Dr. Olivia Morad

This document summarizes stochastic queueing models with unlimited servers (M/M/∞) and parallel channels (M/M/C). For the M/M/∞ model, it describes the state transition diagram and derives the Poisson distribution for the number of customers in the system. It shows that the average waiting time is simply the average service time. For the M/M/C model with C parallel servers, it presents the state transition diagram and derives the probability distribution of the number of customers in the system. It expresses the normalization constant P0 in terms of the traffic intensity ρ.

Uploaded by

mod nod
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
70 views15 pages

Stochastic Models: Dr. Olivia Morad

This document summarizes stochastic queueing models with unlimited servers (M/M/∞) and parallel channels (M/M/C). For the M/M/∞ model, it describes the state transition diagram and derives the Poisson distribution for the number of customers in the system. It shows that the average waiting time is simply the average service time. For the M/M/C model with C parallel servers, it presents the state transition diagram and derives the probability distribution of the number of customers in the system. It expresses the normalization constant P0 in terms of the traffic intensity ρ.

Uploaded by

mod nod
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 15

Stochastic Models

Lecture 8

by
Dr. Olivia Morad
2018
A- Unlimited service M/M/∞
(Responsive servers/Infinite Number of servers)
• We treat a queuing model for which there is unlimited service, that is an
infinite number of servers available.

• This model is often referred to as the "Ample– server problem".


• A self service type situation is one example of such a model.
• We consider a case that may be interpreted either as that of a responsive
server who accelerates his service rate linearly when more customers are
waiting / or may be interpreted as the case where there is always a new
clerk or server available for each arriving customer.
λ λ λ λ

0 1 2 n-1 n n+1

µ 2µ nµ (n+1)µ

"State– transition rate diagram for the infinite – server case M/M/∞"

a- Birth – death Coefficients as follows:


λn = λ n = 0, 1, 2, 3,….

µn = n µ n = 1, 2, 3, ….
λ n
Pn = P
n! µ n 0

Ø We find P0 as usual by using the boundary condition


that is:
∞

∑p
n =0
n =1
 ∞ λn  −1
P0 = ∑ n
 n =0 n! µ 
n
∞
1 λ
∑   , (r = λ / u )
n =0 n!  µ 

−r
∞
rn P0 = e
∑
n =0 n!
= e r
r n e− r (λ µ ) e−λ / µ
n
Pn = ⇒ Pn = n>0
n! n!

b-Poisson Distribution

c- The Expected system size, L, is the mean of the Poisson


Distribution and is: λ
L=
µ
W= ?, Lq= ?, Wq= ?
ØThe average waiting time in the system obviously becomes
merely the average service time hence
1
W= → Why?
µ
ØThe waiting time distribution W(t) is indentical to the
service time distribution, namely Exponential with mean:
1
µ
B- Queues with Parallel Channels M/M/C
• Multi server model in which server has an independently and
identical distributed Exponential service time distribution, with the
arrival process again assumed to be Poisson.

• We Consider the case in which there are C servers.


a- Birth – death Coefficients as follows:
λn = λ for all n
nµ 1 ≤ n < C….(1)
Cµ n ≥ C….. (2)
(1) …When there are fewer than C customers in the system say n < C
only n of C servers are busy and the system is putting out at a
mean rate [n µ].
(2) …. If there are more then C customer all C servers are busy, each
at mean rate µ, mean system output rate [Cµ].

λ λ λ λ λ

0 1 2 C-2 C-1 C

µ 2µ (C-1) µ Cµ
CU

"State – transition rate diagram for M/M/C"


Computational Problem: Book
Page
n −1
λ
n
λ 1 λn
Pn = P0 ∏ = P0   Pn = P 1≤ n≤ C
i = 0 (i + 1) µ  µ  n! n! µ n 0

C −1
λ n −1
λ λn
Pn = P0 ∏ ∏ p n = n −C
i =0 (i + 1) µ j =C C µ C C! µ
P
n 0 n≥ C

ØIn order to find P0, me must again use the boundary


condition
∞

∑p
n =0
n =1
Which gives:

 C −1 λn ∞
λn 
P0 ∑ + ∑ n −C n
=1
 n =0 n! µ C! µ 
n
n =C C

λ
r= → substitute
µ

 C −1 r n ∞ rn 
P0 = ∑ + ∑ n −C  = 1
 n =0 n! n =C C C! 
∞ ∞ n −C
r
n C
r r
∑
n =C
n −C
=
C C! C!
∑
n =C
 
C 

∞ m
r
C
r
= ∑   n– C=m
C! m=0 C 

rC 1
=
C! (1 − r / C )
−1
 C −1
r n
Cr C

P0 =  ∑ + 
 n =0 n! C ! (C − r )
−1
 C −1 1  λ 
n
1 λ
C
 Cµ 
P0 = ∑   +    
 n =0 n!  µ  C!  µ   Cµ − λ 

λ
ρ= <1 L= ??
Cµ
THANK YOU

You might also like