Professional Documents
Culture Documents
=
= =
3
4
15
10
5
10
ALQ = =
=
Result obtained using simulation
program
Exponential arrival and service time is used
r=rand()/32768
Iat=(-1./mue)*log(1-r)
Time is counted in minutes
For a single run
Number of arrivals=50
Average waiting time=30.83minutes
Average server idle time=2.46
Maximum queue length=16
Single queue multiple servers
M/M/s/
Let
s denote number of servers in system
Each server provides service at the same rate
Average arrival rate for all n customers is same
<s
M/M/s/ analysis cont
For n busy servers, the over all service rate is
n
Probability that there are (n+1) customers is
given by( n>s)
Where
(n-s) customers are waiting
M/M/s/ analysis cont
Probability that all servers are busy is
probability that n> s
This is given by
M/M/s/ cont
Average length of queue
=
Example 2 server M/M/2/
In a service station with two servers,
customers arrive at an average rate of 10 per
hour. The service rate of each server is 6
customers/hour.
Determine
A) the fraction of time that all servers are busy
B) average number of customers waiting
C) average waiting time
Soln.
Servers will be busy if there are n>2 customers
P(n>2 )is then
where Po is
Then P(n >2)
2 s
hr / 6
hr / 10
=
=
=
= = 0.79
Analysis using simulation M/M/2/3
2 servers
Maximum capacity of 3
Exponential arrival and service time
0 9 13 22 26 33
10 7 12 20 15 15
arrival Server 1 Server 2
cust idle service wait idle service wait
0 0 10 0
9 - - - 9 7 0
13 3 12 0 - - -
22 - - - 6 20 0
26 1 15 0 - - -
Exercise
Write a simulation program to analyze an
M/D/2/3 system
Exponential arrival with mean 3 minutes
2 servers and maximum capacity of 3
Service time is deterministic with 5 and 7 minute
service time respectively
Simulate system for 1 hour and determine
Idle time of servers
Waiting time of customers