S Queueing Theory. M/G/1-LIFO-PR queue. Samuli Aalto TKK Helsinki University of Technology

Similar documents
S Queueing Theory. M/G/1-PS queue. Samuli Aalto TKK Helsinki University of Technology. Lect_MG1-PS.ppt S Queueing Theory Fall 2009

Of Buses and Bunching:

EE 261 The Fourier Transform and its Applications Fall 2007 Problem Set Two Due Wednesday, October 10

Appendix Y: Queuing Models and Applications

Scalability of delays in input queued switches

MATH 214 (NOTES) Math 214 Al Nosedal. Department of Mathematics Indiana University of Pennsylvania. MATH 214 (NOTES) p. 1/11

Measurements of low energy e e hadronic cross sections and implications for the muon g-2

1.1 The Language of Mathematics Expressions versus Sentences

Linear mixed models and when implied assumptions not appropriate

ECE302H1S Probability and Applications (Updated January 10, 2017)

Math Assignment 10

Sensors, Measurement systems Signal processing and Inverse problems Exercises

Senior Math Studies Lesson Planning Date Lesson Events

Tape. Tape head. Control Unit. Executes a finite set of instructions


Chapter 27. Inferences for Regression. Remembering Regression. An Example: Body Fat and Waist Size. Remembering Regression (cont.)

Measurement of overtone frequencies of a toy piano and perception of its pitch

MID-TERM EXAMINATION IN DATA MODELS AND DECISION MAKING 22:960:575

Sequencing. Lan-Da Van ( 范倫達 ), Ph. D. Department of Computer Science National Chiao Tung University Taiwan, R.O.C. Fall,

A DISCRETE FILTER BANK APPROACH TO AUDIO TO SCORE MATCHING FOR POLYPHONIC MUSIC

Randomness for Ergodic Measures

GS1881, GS4881, GS4981 Monolithic Video Sync Separators

CS8803: Advanced Digital Design for Embedded Hardware

Optimum Frame Synchronization for Preamble-less Packet Transmission of Turbo Codes

Moving on from MSTAT. March The University of Reading Statistical Services Centre Biometrics Advisory and Support Service to DFID


CS8803: Advanced Digital Design for Embedded Hardware

Device/PLC Connection Manuals

More design examples, state assignment and reduction. Page 1

RATE-REDUCTION TRANSCODING DESIGN FOR WIRELESS VIDEO STREAMING

TITLE MANUSCRIPT PREPARATION INSTRUCTIONS. Sydney, NSW 2052, Australia 2 Department of Mechanical Engineering, Auburn University

Problem Weight Total 100

Sampler Overview. Statistical Demonstration Software Copyright 2007 by Clifford H. Wagner

Reconstruction of Ca 2+ dynamics from low frame rate Ca 2+ imaging data CS229 final project. Submitted by: Limor Bursztyn

Digital Delay / Pulse Generator DG535 Digital delay and pulse generator (4-channel)

Meeting Real-Time Constraint of Spectrum Management in TV Black-Space Access

Technical report on validation of error models for n.

ECE 555 DESIGN PROJECT Introduction and Phase 1

2.810 Manufacturing Processes and Systems Quiz #2. November 15, minutes

Research Article Sufficient Conditions and Applications for Carathéodory Functions

10.1 Sequential logic circuits are a type of logic circuit where the output of the circuit depends not only on

Influence of Available Bandwidth on the Statistical Characterization of Compressed Video

ALGEBRAIC PURE TONE COMPOSITIONS CONSTRUCTED VIA SIMILARITY

RGB Encoder For the availability of this product, please contact the sales office. VIDEO OUT Y/C MIX DELAY CLAMP

Hidden Markov Model based dance recognition

UNIVERSITY OF BOLTON SCHOOL OF ENGINEERING BENG (HONS) ELECTRICAL AND ELECTRONIC ENGINEERING SEMESTER 2 EXAMINATION 2016/2017

Week 14 Music Understanding and Classification

FPGA DESIGN OF CLUTTER GENERATOR FOR RADAR TESTING

ECGR 6264 RF Design Midterm Spring 2005

CVOUT Vcc2 TRAP SWITCH Y/C MIX INTERNAL TRAP DELAY LPF LPF SIN-PULSE NPIN SCIN

A Discrete Time Markov Chain Model for High Throughput Bidirectional Fano Decoders

Mixed Models Lecture Notes By Dr. Hanford page 151 More Statistics& SAS Tutorial at Type 3 Tests of Fixed Effects

UNIVERSITY OF HONG KONG Faculty of Engineering M.Sc.(Eng) in Building Services MEBS 6002 Lighting Engineering

Time Domain Simulations

Signal Stability Analyser

Chapter 12. Synchronous Circuits. Contents

Simultaneous Experimentation With More Than 2 Projects

Problems of high DFE coefficients

Classification of Timbre Similarity

BER margin of COM 3dB

Formulate Designer Series 20 Backwall - Kit 07

THE RELATIONSHIP OF BURR HEIGHT AND BLANKING FORCE WITH CLEARANCE IN THE BLANKING PROCESS OF AA5754 ALUMINIUM ALLOY

BOOKLET. Preparing Papers for 15th REAAA Conference in Bali Guidelines for Authors

Duobinary Transmission over ATCA Backplanes

Product datasheet. PARATHOM R50 LED reflector lamps R50 with retrofit screw base. Areas of application. Product benefits.

DISTORTION-AWARE RETRANSMISSION OF VIDEO PACKETS AND ERROR CONCEALMENT USING THUMBNAIL. Zhi Li. EE398 Course Project, Winter 07/08

Comment #147, #169: Problems of high DFE coefficients

Quarter wave retarders for Dense Wave Division Multiplexing ABSTARCT 1. INTRODUCTION

Magnetic Sensor - Incremental EHP

NH 67, Karur Trichy Highways, Puliyur C.F, Karur District UNIT-III SEQUENTIAL CIRCUITS

Research Article Design and Analysis of a High Secure Video Encryption Algorithm with Integrated Compression and Denoising Block

arxiv: v2 [cs.ai] 3 Aug 2016

Formulate Designer Series 10 Backwall - Kit 05

Course Web site:

Department of Electrical & Electronic Engineering Imperial College of Science, Technology and Medicine. Project: Real-Time Speech Enhancement

Formulate Designer Series 30 Backwall - Kit 05

Lecture 11: Synchronous Sequential Logic

11.1 As mentioned in Experiment 10, sequential logic circuits are a type of logic circuit where the output

CHAPTER 3 SEPARATION OF CONDUCTED EMI

A New Method for Processing Impact Excited Continuous-Scan Laser. Doppler Vibrometer Measurements

Likelihood inference for Archimedean copulas

Refined Spectral Template Models for Score Following

1 Introduction to the life course perspective. 2 Working with life course data. 3 Familial life course analysis. 4 Visualization.

Design Project: Designing a Viterbi Decoder (PART I)

Design of NIR Chromenylium-Cyanine Fluorophore Library for Switch-ON and Ratiometric Detection of Bio-Active Species in Vivo

with - < n < +. There are two types of approximations associated with the sampling process. finite precision of the ADC finite sampling frequency.

Available online at ScienceDirect. Procedia Computer Science 46 (2015 )

Power (dbm) λ (nm) LINK DISTANCE SDI Bit Rate Max. Link Distance (km) 3G-SDI 2.97Gbps 30 HD-SDI 1.485Gbps 30 SD-SDI 270Mbps 30

HCF4054B 4 SEGMENT LIQUID CRYSTAL DISPLAY DRIVER WITH STROBED LATCH FUNCTION

Jazz Melody Generation and Recognition

About this Manual. Support for Your Product

Features of the 745T-20C: Applications of the 745T-20C: Model 745T-20C 20 Channel Digital Delay Generator

Rocky Mountain Population Sandhill Cranes

Digital Logic Design I

Chapter 6. Flip-Flops and Simple Flip-Flop Applications

Predictability of Music Descriptor Time Series and its Application to Cover Song Detection

EXPERIMENTS WITH THE ACTIVE VIBRATION CONTROL OF A CANTILEVER BEAM

GLM Example: One-Way Analysis of Covariance

Contents. Iannis Xenakis s Achorripsis (1957) : The Matrix Game. Iannis Xenakis ( ) Iannis Xenakis Achorripsis (1957) Distribution of Events


Transcription:

S-38.3143 Queueing Theory M/G/1-LIFO-PR queue Samuli Aalto TKK Helsinki University of Technology Lect_MG1-LIFO-PR.ppt S-38.3143 Queueing Theory Fall 2009 1

Contents Preliminaries Queue length distribution for M/M/1-LIFO-PR Queue length distribution for M/E K /1-LIFO-PR Queue length distribution for M/G/1-LIFO-PR Unfinished work distribution for M/E K /1-LIFO-PR Unfinished work distribution for M/G/1-LIFO-PR Busy period distribution for M/G/1-LIFO-PR 2

M/G/1-LIFO-PR queue Customers arrive according to a Poisson process at rate λ IID inter-arrival times exponential inter-arrival time distribution with mean 1/λ Customers are served by 1 server according to the LIFO-PR service discipline IID service times a general service time distribution with mean 1/µ There are customer places in the system λ µ 3

Service disciplines LIFO-PR Preemptive Last In First Out (LIFO-PR) customers are served one-by-one but not necessarily until completion when a new customer arrives, the current service is stopped and the new customer is taken into service when a server becomes free, the last arrived customer is taken into service thus, the customers in the system constitute a stack, and the system serves always the customer that has been waiting for the shortest time preemptive, work-conserving, and non-anticipating discipline also known as preemptive Last Come First Served (LCFS-PR) 4

Preliminaries (1) 5

Preliminaries (2) 6

Preliminaries (3) 7

Contents Preliminaries Queue length distribution for M/M/1-LIFO-PR Queue length distribution for M/E K /1-LIFO-PR Queue length distribution for M/G/1-LIFO-PR Unfinished work distribution for M/E K /1-LIFO-PR Unfinished work distribution for M/G/1-LIFO-PR Busy period distribution for M/G/1-LIFO-PR 8

M/M/1-LIFO-PR queue Customers arrive according to a Poisson process at rate λ IID inter-arrival times exponential inter-arrival time distribution with mean 1/λ Customers are served by 1 server according to the LIFO-PR service discipline IID service times exponential service time distribution with mean 1/µ There are customer places in the system λ µ 9

Exponential service time distribution S Exp ( µ ), µ > 0 P{S (x,x+h] S > x} = µh + o(h), where o(h)/h 0 as h 0 Value space: (0, ) µ PDF and CDF: f ( x) = µ e µ x F( x) = 1 e µ x 10

Moments and the Laplace transform S Exp ( µ ), µ > 0 Mean value: E[S] = 0 µx e -µx dx = 1/µ Second moment: E[S 2 ] = 0 µx 2 e -µx dx = 2/µ 2 Variance: D 2 [S] = E[S 2 ] E[S] 2 = 1/µ 2 Standard deviation: D[S] = D 2 [S] = 1/µ Coefficient of variation: C[S] = D[S]/E[S] = 1 Laplace transform: E[e -ss ] = 0 µ e -(µ+s)x dx = µ/(µ+s) 11

State transition diagram of X(t) Let X(t) denote the number of customers in the system at time t Assume that X(t) = i at some time t, and consider what happens during a short time interval (t, t+h]: with prob. λh + o(h), a new customer arrives (state transition i i+1) if i > 0, then, with prob. µh + o(h) = µh + o(h), a customer leaves the system (state transition i i 1) Process X(t) is clearly a Markov process with state transition diagram 0 λ µ λ 1 µ 2 Note that process X(t) is an irreducible birth-death process with an infinite state space S = {0,1,2,...} λ µ 12

13 M/G/1-LIFO-PR queue Equilibrium distribution Detailed balance equations (DBE): Normalizing condition (N): 0,1,2,K, (DBE) 0 1 1 = = = = = + + i i i i i i i i π ρ π ρπ π π µ π λ π µ λ (N) 1 0 0 0 = = = = i i i i ρ π π ( ) 1 if, 1 1 1 1 1 0 0 < = = = = ρ ρ ρ π ρ i i

Queue length distribution for M/M/1-LIFO-PR (1) 14

Queue length distribution for M/M/1-LIFO-PR (2) 15

Queue length distribution for M/M/1-LIFO-PR (3) 16

Queue length distribution for M/M/1-LIFO-PR (4) 17

Contents Preliminaries Queue length distribution for M/M/1-LIFO-PR Queue length distribution for M/E K /1-LIFO-PR Queue length distribution for M/G/1-LIFO-PR Unfinished work distribution for M/E K /1-LIFO-PR Unfinished work distribution for M/G/1-LIFO-PR Busy period distribution for M/G/1-LIFO-PR 18

M/E K /1-PS queue Customers arrive according to a Poisson process at rate λ IID inter-arrival times exponential inter-arrival time distribution with mean 1/λ Customers are served by 1 server according to the PS service discipline IID service times Erl(K, Kµ) service time distribution with K phases and mean 1/µ There are customer places in the system λ µ 19

Erlang service time distribution S Erl ( K, Kµ ), µ > 0 IID exponential phases in a series; S = S 1 + + S K where S j Exp(Kµ) K = total number of phases µ = intensity of any single phase Value space: (0, ) PDF and CDF: f ( x) = F( x) Kµ = 1 K 1 ( Kµ x) ( K 1)! e K ( K x K µ ) ( K k)! k= 1 Kµ x k Kµ Kµ Kµ 1 2 K e Kµ x 20

Moments and the Laplace transform S Erl ( K, Kµ ), µ > 0 Mean value: E[S] = E[S 1 ] + + E[S K ] = K/(Kµ) = 1/µ Variance: D 2 [S] = D 2 [S 1 ] + + D 2 [S K ] = K/(Kµ) 2 = 1/(Kµ 2 ) Second moment: E[S 2 ] = E[S] 2 + D 2 [S] = (1+1/K)/µ 2 Standard deviation: D[S] = D 2 [S] = 1/(( K)µ) Coefficient of variation: C[S] = D[S]/E[S] = 1/( K) 1 Laplace transform : E[e -ss ] = E[e -ss 1 ] E[e -ss K ] = (Kµ/(Kµ+s)) K Kµ Kµ Kµ 1 2 K 21

Queue length distribution for M/E K /1-LIFO-PR (1) Kµ Kµ Kµ 1 2 K 22

Queue length distribution for M/E K /1-LIFO-PR (2) 1 2 K 23

Queue length distribution for M/E K /1-LIFO-PR (3) 1 2 K 24

Queue length distribution for M/E K /1-LIFO-PR (4) 1 2 K 25

Queue length distribution for M/E K /1-LIFO-PR (5) 26

Queue length distribution for M/E K /1-LIFO-PR (6) 27

Queue length distribution for M/E K /1-LIFO-PR (7) 28

Queue length distribution for M/E K /1-LIFO-PR (8) 29

Queue length distribution for M/E K /1-LIFO-PR (9) 1 2 K 30

Queue length distribution for M/E K /1-LIFO-PR (10) 1 2 K 31

Queue length distribution for M/E K /1-LIFO-PR (11) 1 2 K 32

Queue length distribution for M/E K /1-LIFO-PR (12) 33

Queue length distribution for M/E K /1-LIFO-PR (13) 34

Queue length distribution for M/E K /1-LIFO-PR (14) 35

Contents Preliminaries Queue length distribution for M/M/1-LIFO-PR Queue length distribution for M/E K /1-LIFO-PR Queue length distribution for M/G/1-LIFO-PR Unfinished work distribution for M/E K /1-LIFO-PR Unfinished work distribution for M/G/1-LIFO-PR Busy period distribution for M/G/1-LIFO-PR 36

Queue length distribution for M/G/1-LIFO-PR (1) 37

Queue length distribution for M/G/1-LIFO-PR (2) 38

Queue length distribution for M/G/1-LIFO-PR (3) 39

Queue length distribution for M/G/1-LIFO-PR (4) 40

Queue length distribution for M/G/1-LIFO-PR (5) 41

Queue length distribution for M/G/1-LIFO-PR (6) 42

Queue length distribution for M/G/1-LIFO-PR (7) 43

Queue length distribution for M/G/1-LIFO-PR (8) 44

Queue length distribution for M/G/1-LIFO-PR (9) 45

Contents Preliminaries Queue length distribution for M/M/1-LIFO-PR Queue length distribution for M/E K /1-LIFO-PR Queue length distribution for M/G/1-LIFO-PR Unfinished work distribution for M/E K /1-LIFO-PR Unfinished work distribution for M/G/1-LIFO-PR Busy period distribution for M/G/1-LIFO-PR 46

Unfinished work distribution for M/E K /1-LIFO-PR (1) Kµ Kµ Kµ 1 2 K 47

Unfinished work distribution for M/E K /1-LIFO-PR (2) 48

Unfinished work distribution for M/E K /1-LIFO-PR (3) 49

Unfinished work distribution for M/E K /1-LIFO-PR (4) 50

Unfinished work distribution for M/E K /1-LIFO-PR (5) 51

Unfinished work distribution for M/E K /1-LIFO-PR (6) 52

Unfinished work distribution for M/E K /1-LIFO-PR (7) 53

Unfinished work distribution for M/E K /1-LIFO-PR (8) 54

Unfinished work distribution for M/E K /1-LIFO-PR (9) 55

Unfinished work distribution for M/E K /1-LIFO-PR (10) 56

Unfinished work distribution for M/E K /1-LIFO-PR (11) 57

Unfinished work distribution for M/E K /1-LIFO-PR (12) 58

Unfinished work distribution for M/E K /1-LIFO-PR (13) 59

Unfinished work distribution for M/E K /1-LIFO-PR (14) 60

Unfinished work distribution for M/E K /1-LIFO-PR (15) 61

Contents Preliminaries Queue length distribution for M/M/1-LIFO-PR Queue length distribution for M/E K /1-LIFO-PR Queue length distribution for M/G/1-LIFO-PR Unfinished work distribution for M/E K /1-LIFO-PR Unfinished work distribution for M/G/1-LIFO-PR Busy period distribution for M/G/1-LIFO-PR 62

Unfinished work distribution for M/G/1-LIFO-PR (1) 63

Unfinished work distribution for M/G/1-LIFO-PR (2) 64

Unfinished work distribution for M/G/1-LIFO-PR (3) 65

Unfinished work distribution for M/G/1-LIFO-PR (4) 66

Unfinished work distribution for M/G/1-LIFO-PR (5) 67

Unfinished work distribution for M/G/1-LIFO-PR (6) 68

Unfinished work distribution for M/G/1-LIFO-PR (7) 69

Contents Preliminaries Queue length distribution for M/M/1-LIFO-PR Queue length distribution for M/E K /1-LIFO-PR Queue length distribution for M/G/1-LIFO-PR Unfinished work distribution for M/E K /1-LIFO-PR Unfinished work distribution for M/G/1-LIFO-PR Busy period distribution for M/G/1-LIFO-PR 70

Busy period distribution for M/G/1-LIFO-PR (1) 71

Busy period distribution for M/G/1-LIFO-PR (2) 72

Busy period distribution for M/G/1-LIFO-PR (3) 73

Busy period distribution for M/G/1-LIFO-PR (4) 74

Busy period distribution for M/G/1-LIFO-PR (5) 75

Busy period distribution for M/G/1-LIFO-PR (6) 76

Busy period distribution for M/G/1-LIFO-PR (7) 77

Busy period distribution for M/G/1-LIFO-PR (8) 78

Busy period distribution for M/G/1-LIFO-PR (9) 79

Busy period distribution for M/G/1-LIFO-PR (10) 80

Busy period distribution for M/G/1-LIFO-PR (11) 81

Busy period distribution for M/G/1-LIFO-PR (12) 82

Busy period distribution for M/G/1-LIFO-PR (13) 83

Busy period distribution for M/G/1-LIFO-PR (14) 84

Busy period distribution for M/G/1-LIFO-PR (15) 85

Busy period distribution for M/G/1-LIFO-PR (16) 86

References 87