Upload
nikhil-malhotra
View
12
Download
0
Tags:
Embed Size (px)
DESCRIPTION
Waiting line model
Citation preview
Advance Waiting Line Theory and
Simulation Modeling
© 2008 Pearson Prentice Hall --- Introduction to Operations and Supply Chain Management, 2/e --- Bozarth and Handfield, ISBN: 0131791036
Supplement 8S, Slide 2
Supplement Objectives
Be able to: Describe different types of waiting line systems. Use statistics-based formulas to estimate waiting
line lengths and waiting times for three different types of waiting line systems.
Explain the purpose, advantages and disadvantages, and steps of simulation modeling.
Develop a simple Monte Carlo simulation using Microsoft Excel.
Develop and analyze a system using SimQuick.
© 2008 Pearson Prentice Hall --- Introduction to Operations and Supply Chain Management, 2/e --- Bozarth and Handfield, ISBN: 0131791036
Supplement 8S, Slide 3
Alternative Waiting Lines
• Single-Channel, Single-Phase– Ticket window at theater,
• Multiple-Channel, Single-Phase– Tellers at the bank, windows at post office
• Single-Channel, Multiple-Phase– Line at the Laundromat, DMV
© 2008 Pearson Prentice Hall --- Introduction to Operations and Supply Chain Management, 2/e --- Bozarth and Handfield, ISBN: 0131791036
Supplement 8S, Slide 4
Alternative Waiting Lines
Single-Channel, Single-Phase
Multiple-Channel, Single-Phase
Single-Channel, Multiple-Phase
© 2008 Pearson Prentice Hall --- Introduction to Operations and Supply Chain Management, 2/e --- Bozarth and Handfield, ISBN: 0131791036
Supplement 8S, Slide 5
Assumptions
• Arrivals– At random (Poisson, exponential distributions)– Fixed (appointments, service intervals)
• Service times– Variable (exponential, normal distributions)– Fixed (constant service time)
• Other– Size of arrival population, priority rules,
balking, reneging
© 2008 Pearson Prentice Hall --- Introduction to Operations and Supply Chain Management, 2/e --- Bozarth and Handfield, ISBN: 0131791036
Supplement 8S, Slide 6
Poisson Distribution
Probability of n arrivals in T time periods
where = arrival rate
Tn
n enT
P !
)(
© 2008 Pearson Prentice Hall --- Introduction to Operations and Supply Chain Management, 2/e --- Bozarth and Handfield, ISBN: 0131791036
Supplement 8S, Slide 7
Waiting Line Formulas
© 2008 Pearson Prentice Hall --- Introduction to Operations and Supply Chain Management, 2/e --- Bozarth and Handfield, ISBN: 0131791036
Supplement 8S, Slide 8
P0 = Probability of 0 Units in Multiple-Channel System
MM
Mn
PM
M
n
n
!1
!1
1
1
0
0
© 2008 Pearson Prentice Hall --- Introduction to Operations and Supply Chain Management, 2/e --- Bozarth and Handfield, ISBN: 0131791036
Supplement 8S, Slide 9
Single-Channel, Single-PhaseManual Car Wash Example
• Arrival rate = 7.5 cars per hour• Service rate = an average of 10 cars per hour• Utilization = / = 75%
© 2008 Pearson Prentice Hall --- Introduction to Operations and Supply Chain Management, 2/e --- Bozarth and Handfield, ISBN: 0131791036
Supplement 8S, Slide 10
Single-Channel, Single-PhaseAutomated Car Wash Example
• Arrival rate = 7.5 cars per hour• Service rate = a constant rate of 10 cars per hour• Utilization = / = 75%
© 2008 Pearson Prentice Hall --- Introduction to Operations and Supply Chain Management, 2/e --- Bozarth and Handfield, ISBN: 0131791036
Supplement 8S, Slide 11
Comparisons
Manual wash, single server
Automated wash, single server
Manual wash, two servers
Cars waiting
2.25 1.125 0.1227
Cars in system
3 1.875 1.517
Time waiting
18 minutes 9 minutes 1 minute
Time in System
24 minutes 15 minutes 7 minutes
© 2008 Pearson Prentice Hall --- Introduction to Operations and Supply Chain Management, 2/e --- Bozarth and Handfield, ISBN: 0131791036
Supplement 8S, Slide 12
Simulation Modeling
Advantages• Off-line evaluation of
new processes or process changes
• Time compression• “What-if” analysis• Provides variance
estimates in addition to averages
Disadvantages• Does not provide
optimal solution• More realistic the
more costly and more difficult to interpret
• Still just a simulation
© 2008 Pearson Prentice Hall --- Introduction to Operations and Supply Chain Management, 2/e --- Bozarth and Handfield, ISBN: 0131791036
Supplement 8S, Slide 13
Monte Carlo Simulation
• Maps random numbers to cumulative probability distributions of variables
• Probability distributions can be either discrete (coin flip, roll of a die) or continuous (exponential service time or time between arrivals)
• Random numbers 0 to 99 supplied by computer functions such as = INT(100*RAND()) in Excel.
© 2008 Pearson Prentice Hall --- Introduction to Operations and Supply Chain Management, 2/e --- Bozarth and Handfield, ISBN: 0131791036
Supplement 8S, Slide 14
Monte Carlo Simulation Examples
• Coin toss: Random numbers 0 to 49 for ‘heads’, 50 to 99 for ‘tails’
• Dice throw: Use Excel function = RANDBETWEEN(1,6) for throws
• Service time: Use Excel function = –(avg service time)*ln(RAND()) for exponential service time
© 2008 Pearson Prentice Hall --- Introduction to Operations and Supply Chain Management, 2/e --- Bozarth and Handfield, ISBN: 0131791036
Supplement 8S, Slide 15
Building a Simulation Model
Four basic steps1) Develop a picture of system to be modeled (process
mapping)2) Identify objects, elements, and probability distributions
that define the system Objects = items moving through system Elements = pieces of the system
3) Determine experiment conditions (constraints) and desired outputs
4) Build and test model, capture the output data
© 2008 Pearson Prentice Hall --- Introduction to Operations and Supply Chain Management, 2/e --- Bozarth and Handfield, ISBN: 0131791036
Supplement 8S, Slide 16
Simulation Example(Using single-channel, single-phase waiting line)
1) Process map
2) Time between arrivals (exponential distribution), service time (exponential distribution), objects = cars, elements = line and wash station
3) Maximum length for line, time spent in the system4) Run model for a total of 100 cars entering the car
wash, average the results for waiting time, cars in line, etc.
‘SimQuick’ Simulation
An Excel-based application for simulating processes that allows use of constraints
(see text example 8S.5)