Está en la página 1de 6

models

How to choose a queueing model


All models in this workbook are Poisson arrivals, infinite population, and FCFS.
The models differ by
(1) the service time distribution (exponential, constant or general)
(2) the number of servers (single server or multiple servers)
(3) waiting room capacity (unlimited waiting room or limited waiting room buffer)

M/M/1

Service time distribution Exponential


Number of servers
Single
Waiting room capacity
Unlimited

M/D/1

M/G/1

Constant
Single
Unlimited

General
Single
Unlimited

The Erlang Loss Model is a special case of M/M/s/b model where the waiting room capacity (b) is zero.
The following models assume exponential distribution of service time:
M/M/1, M/M/s, M/M/s/b
The following models assume constant service time:
M/C/1
The following models assume generally distributed (such as normal) service time:
M/G/1
The following models assume a single server:
M/M/1, M/C/1, M/G/1
The following models assume multiple servers:
M/M/s, M/M/s/b
The following models assume unlimited waiting room capacity:
M/M/1, M/C/1, M/G/1, M/M/s
The following models assume limited waiting room capacity:
M/M/s/b

Page 1

models

M/M/s

M/M/s/b

Exponential
Multiple
Unlimited

Exponential
Multiple
Limited

m capacity (b) is zero.

Page 2

M-M-1
Model 1 (M/M/1 Queue):

Inputs

Outputs

Summary measures

Page 3

M-M-1
Model 1 (M/M/1 Queue):
Single server, Infinite population, Poisson arrivals, FCFS, Exponential service time, Unlimited queue length
Yellow cells need user inputed values

Inputs
Units of time
Arrival rate (lambda)
Service rate (mu)

Outputs
Direct outputs from inputs
Mean time between arrivals
Mean time per service
Traffic intensity
Summary measures
Utilization rate of server OR Prob that arriving unit has to wait for service
Average number of customers waiting in line (Lq)
Average number of customers in system (L)
Average time waiting in line (Wq)
Average time in system (W)
Probability of no customers in system (P0)
Distribution of number of customers in system
n (customers)
6
Distribution of time in queue
t (time in queue)
0.25

Page 4

M-M-1

hour
8 cusomters per
10 cusomters per

hour
hour

0.125 hour
0.1 hour
0.8

0.80
3.2
4.00
0.4
0.5
0.20

customers
customers
hour
hour
(this is the probability of empty system)

P(n in system)
0.052429

P(wait > t)
0.485225

Page 5

Model 5: Multiple servers with finite buffers

Model 5 (M/M/s/b queue):


Multiple servers, Infinite population, Poisson arrival, FCFS, Exponential service time, Limited buffer (b)
Yellow cells need user inputed values

Inputs
Unit of time
Arrival rate (lambda)
Service rate (mu)
Number of identical servers (s)
Buffer (waiting room) size

hour
0.75
1
2
50

customers per
customers per
servers
customers

hour
hour
(The maximum number in system is equal to number of servers plus waiting room size)

Outputs
Direct outputs from inputs
Mean time between arrivals
Mean time per service
Traffic intensity
Summary measures
Average utilization rate of servers
Average number of customers waiting in line (Lq)
Average number of customers in system (L)
Average time waiting in line (Wq)
Average time in system (W)
Probability of no customers in system (P0)
Probability of rejecting a customer (balking rate)
Effective arrival rate

1.333 hour
1 hour
0.375

#VALUE!
#VALUE!
#VALUE!
#VALUE!
#VALUE!
#VALUE!
#VALUE!
#VALUE!

customers
customers
hour
hour
(Probability of empty system)
(Reject rate)
(Entering rate)

También podría gustarte