Operations Research

OPERATIONS RESEARCH QUEUEING THEORY MULTIPLE SERVER MODELS
MULTIPLE-SERVER QUEUE Number of service channels is more than one Frequently encountered in telecommunications or a customer service environment. Here, care is needed to ensure that it is a multiple servers queue, not a network of single server queues, because results may differ depending on how the queuing model behaves. A single queue with multiple servers performs better than each server having their own queue A single large pool of servers performs better than two or more smaller pools, even though there are the same total number of servers in the system.
MULTIPLE SERVER SINGLE STAGE QUEUE
Customers In queue
Service Facilities
Supermarkets have many cash counters opened. Gasoline stations have many pumps opened and many attendants to be on duty. Manufacturing plants determine the optimal number of mechanics on duty in each shift to repair machines that break down. Banks may have multiple teller windows to keep open to serve customers during peak hours. Air ticket sales Toll Booths on National Highways Air Traffic Control : Opening of a new runway at times of bad weather. Elevators at malls
EXAMPLES
The M/M/c:GD// Queuing System

Poisson arrivals with rate Exponential service times with parameter c parallel servers There is a single line of customers waiting to be served at one of the c parallel servers Arriving customer finds n customers in system n < c: it is routed to any idle server n c: it joins the waiting queue all servers are busy
n= n=0,, n=n n=1,,c-1 n=c n=c,,
TRANSITION RATE DIAGRAM

2
2 3
c
(c-1)
c+1 c+2
c c c
Using the equality rate principle, we have the balance equations -(n +n)Pn+n+1P n+1+ n-1 P n-1=0; -0 P0 +1 P1 =0 ---- (*) =>P1= 0 P0/ 1 Further, -(1 +1)P1+2P 2+ 0 P 0=0 P2= =>P2= 1 P1/ 2= 0 1 P0/ 1 2 Thus in general, Pn= (using (*))
In our case, we have Pn= nP0/(2 )(3 )(n )= P0 for n<c When nc Pn= nP0/(c )(c )(c )(c )(c-1) 2 1 = P0 Thus if = /, the traffic intensity, we have Pn = Po n<c Po nc
To solve for Po We have

=>P0=
PERFORMANCE MEASURES
Average number Lq of customers in the queue: Lq = Average number Ls of customers in the system: Ls= Ls = Lq + Average Time a customer spends in the system: Ws = (1/) Ls Average Time a customer spends in the Queue: Wq = (1/) Lq
To solve for Lq Lq = Put n-c=k Lq= = =

= =
MODEL
Students arriving at a student registration center of a university must have their registration materials processed by an operator seated at a computer terminal. The system design calls for say 4 operators to be on duty, each operator performing an identical service. Students arrive according to a Poisson process at an average rate of say 100 per hour. Each operator can process 40 students per hour with service time being exponentially distributed.
The Emergency room of COUNTY HOSPITAL provides quick medical care for emergency cases brought to the hospital by ambulance or private automobile. At any hour there is always one doctor on duty in the emergency room. There has been a continuous increase in the number of emergency room visits every year. Patients arriving during peak usage hours (the early evening) have to wait until it is their turn to be treated by the doctor. Therefore, a second doctor is assigned to the emergency room during these hours. The emergency room forms a queuing system. The management engineer has concluded that the emergency cases arrive pretty much at random (a Poisson input process), so that inter arrival times have an exponential distribution. Also time spent by a doctor treating the cases approximately follows an exponential distribution.
COMPARISON OF PERFORMANCE BETWEEN M/M/1:GD// AND M/M/c:GD// QUEUE
Universal Bank has two tellers working on savings account. First Teller handles withdrawals. Second Teller handles deposits. Service time distributions of both deposits and withdrawals are exponential with mean service time 2min per customer. Deposits and withdrawals arrive in Poisson fashion with mean arrival rate 20 per hour.
=20 per hour=1/3 per minute =1/2 ; c=2 Case 1: Treating depositors and withdrawers as units of M/M/1 system Average Waiting time in a queue: Wq =
= = 4minutes
Case 2:If each teller handles both depositor and withdrawals Po = 1/2 Lq =1/12 Wq = (1/) Lq = 1/4 minutes Thus when both teller handles both withdrawal and deposits, the expected waiting time is reduced.
CONCLUSION
Waiting lines are the most frequently encountered problems in every day life. Common to all these cases are the arrivals of objects requiring service and the delays when the service mechanism is busy. Waiting lines cannot be eliminated completely, but waiting time can be reduced. Long waiting time result in loss of customers. Providing additional service facilities reduces waiting time, but may result in an increase in idle time. Optimum model has to be found that minimizes waiting time at minimum investment cost.
REFERENCES
Hamdy A Taha (2002). Operations Research. Prentice Hall of India Frederick S Hillier, Gerald J Lieberman(2004). Introduction to Operations Research. McGraw hill Companies. Ravindra, Phillips, Solberg .Operations Research: Principles and Practice(2nd edition).Wiley India Pvt. Ltd. Donald Gross ,Carl M Harris. Fundamentals of Queuing theory (3rd edition ).Wiley series.

Operations Research

Uploaded by

Copyright:

Available Formats

Operations Research

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Operations Research

Uploaded by

Copyright:

Available Formats

OPERATIONS RESEARCH QUEUEING THEORY MULTIPLE SERVER MODELS

MULTIPLE SERVER SINGLE STAGE QUEUE

The M/M/c:GD// Queuing System

TRANSITION RATE DIAGRAM

To solve for Po We have

To solve for Lq Lq = Put n-c=k Lq= = =

COMPARISON OF PERFORMANCE BETWEEN M/M/1:GD// AND M/M/c:GD// QUEUE

You might also like