# APME_8 - Applied Probability Methods for Engineers Slide...

This preview shows pages 1–10. Sign up to view the full content.

Click to edit Master subtitle style Applied Probability Methods for Engineers Slide Set 8

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
Click to edit Master subtitle style Chapter 20 of Winston’s Operations Research Book Queuing Theory
Queuing Theory n A queue is simply a line n Many processes require that people or items wait for service or processing n We are interested in modeling such processes to answer questions about n How much time is spent in line? n What proportion of time is a server idle? n What is the average number of people/items waiting? n What is the probability distribution of the number of customers or of the waiting time?

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
Queuing Terms n Arrival process: arrivals are called customers n We will assume customers arrive one at a time n Output/service process: we specify a service time distribution n We will assume service time is independent of # of customers in the system n We will consider both parallel and series processors
Queuing Disciplines n Queue discipline gives rule(s) for order of service n First come, first served (FCFS) n Also called first in, first out (FIFO) n Last come, first served (LCFS) (aka LIFO) n Exiting from an elevator n Service in random order (SIRO) n Priority queuing n FCFS within each priority class, where higher priority classes served first

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
Modeling Arrival Process n Let ti be time of the ith arrival n Ti = ti+1 – ti, the ith interarrival time n Assume Ti’s are independent, continuous and identically distributed and described by random variable A n A has density function a(t) n Let 1/λ denote mean interarrival time (measured in hours) n λ is the arrival rate (measured in arrivals per hour) n Under exponential interarrival assumption, E(A) = 1/λ and Var(A) = 1/λ2
Exponential Assumption n Consider P(A > t + h| A ≥ t) = P(A > t + h, A ≥ t)/P(A ≥ t) = P(A > t + h)/P(A ≥ t) = e-λ(t + h)/e-λt = e-λh = P(A > h) n Interarrival times are exponential with parameter λ if and only if the number of arrivals during an interval of length t is Poisson with parameter λt n In general, if arrival rate is stationary, bulk arrivals cannot occur, and past arrivals do not affect future arrivals, then interarrival times are exponential and number of arrivals during time t is Poisson

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
Erlang Distribution n Erlang is a distribution with rate parameter R, shape parameter k (k must be a positive integer), and density function n If T is Erlang distributed with parameters R and k, then E(T) = k/R and Var(T) = k/R2 n When k = 1, Erlang is exponential n Sum of k iid exponential random variables, each with parameter R gives Erlang with parameters k and R ( 29 ( 29 ( 29 1 1 ! k Rt R Rt e f t k - - = -
Modeling Service Process n Service process is defined by a random variable S with density function s(t) and mean service time 1/µ (hours per customer) n Exponential assumption may be a poor assumption for service times n Sometimes use Erlang service times with shape parameter k and rate parameter kµ, which implies a mean service time of 1/µ

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
This is the end of the preview. Sign up to access the rest of the document.

## This note was uploaded on 10/27/2010 for the course ESI 6321 taught by Professor Josephgeunes during the Spring '07 term at University of Florida.

### Page1 / 56

APME_8 - Applied Probability Methods for Engineers Slide...

This preview shows document pages 1 - 10. Sign up to view the full document.

View Full Document
Ask a homework question - tutors are online