M/G/1/GD/ / System. ! Pollaczek-Khinchin (PK) Equation. ! Steady-state probabilities. ! Finding L, W q, W. ! π 0 = 1 ρ

Similar documents
Dennis Bricker, 2001 Dept of Industrial Engineering The University of Iowa. MDP: Taxi page 1

Equilibrium Analysis of the M/G/1 Queue

Remember: Project Proposals are due April 11.

TELCOM 2130 Time Varying Queues. David Tipper Associate Professor Graduate Telecommunications and Networking Program University of Pittsburgh Slides 7

4. Eccentric axial loading, cross-section core

Dynamic Power Management in a Mobile Multimedia System with Guaranteed Quality-of-Service

UNIVERSITY OF IOANNINA DEPARTMENT OF ECONOMICS. M.Sc. in Economics MICROECONOMIC THEORY I. Problem Set II

Queuing system theory

Meenu Gupta, Man Singh & Deepak Gupta

Rank One Update And the Google Matrix by Al Bernstein Signal Science, LLC

DCDM BUSINESS SCHOOL NUMERICAL METHODS (COS 233-8) Solutions to Assignment 3. x f(x)

Applied Statistics Qualifier Examination

Real-Time Systems. Multiprocessor scheduling. Multiprocessor scheduling. Multiprocessor scheduling

The Schur-Cohn Algorithm

Network of Markovian Queues. Lecture

Name: SID: Discussion Session:

Variable time amplitude amplification and quantum algorithms for linear algebra. Andris Ambainis University of Latvia

Principle Component Analysis

Chapter Newton-Raphson Method of Solving a Nonlinear Equation

Jens Siebel (University of Applied Sciences Kaiserslautern) An Interactive Introduction to Complex Numbers

Math 426: Probability Final Exam Practice

First day August 1, Problems and Solutions

INTRODUCTION TO COMPLEX NUMBERS

Using the Econometric Models in Planning the Service of Several Machines at Random Time Intervals. Authors:

GRADE 4. Division WORKSHEETS

TANDEM QUEUE WITH THREE MULTISERVER UNITS AND BULK SERVICE WITH ACCESSIBLE AND NON ACCESSBLE BATCH IN UNIT III WITH VACATION

Lecture 4: November 17, Part 1 Single Buffer Management

Physics 121 Sample Common Exam 2 Rev2 NOTE: ANSWERS ARE ON PAGE 7. Instructions:

= z 20 z n. (k 20) + 4 z k = 4

ψ ij has the eigenvalue

Analysis of Discrete Time Queues (Section 4.6)

Section 6: Area, Volume, and Average Value

Introduction to Continuous-Time Markov Chains and Queueing Theory

Trigonometry. Trigonometry. Solutions. Curriculum Ready ACMMG: 223, 224, 245.

Is there an easy way to find examples of such triples? Why yes! Just look at an ordinary multiplication table to find them!

Dynamic Power Management in a Mobile Multimedia System with Guaranteed Quality-of-Service

Suggested solutions for the exam in SF2863 Systems Engineering. June 12,

Chapter Newton-Raphson Method of Solving a Nonlinear Equation

DIRECT CURRENT CIRCUITS

Chemical Reaction Engineering

Smart Motorways HADECS 3 and what it means for your drivers

The Number of Rows which Equal Certain Row

In this Chapter. Chap. 3 Markov chains and hidden Markov models. Probabilistic Models. Example: CpG Islands

ROB EBY Blinn College Mathematics Department

7.1 Integral as Net Change and 7.2 Areas in the Plane Calculus

Departure Process from a M/M/m/ Queue

7.2 Volume. A cross section is the shape we get when cutting straight through an object.

Chemical Reaction Engineering

Lecture 4: Piecewise Cubic Interpolation

6. Stochastic processes (2)

Graph Theory. Dr. Saad El-Zanati, Faculty Mentor Ryan Bunge Graduate Assistant Illinois State University REU. Graph Theory

6. Stochastic processes (2)

Module 9. Lecture 6. Duality in Assignment Problems

1 Probability Density Functions

Section 3.2: Negative Exponents

The optimal delay of the second test is therefore approximately 210 hours earlier than =2.

AT100 - Introductory Algebra. Section 2.7: Inequalities. x a. x a. x < a

8. INVERSE Z-TRANSFORM

Please review the following statement: I certify that I have not given unauthorized aid nor have I received aid in the completion of this exam.

n α j x j = 0 j=1 has a nontrivial solution. Here A is the n k matrix whose jth column is the vector for all t j=0

Review of Taylor Series. Read Section 1.2

STAT 511 FINAL EXAM NAME Spring 2001

Review of linear algebra. Nuno Vasconcelos UCSD

1 From NFA to regular expression

Section 5.1 #7, 10, 16, 21, 25; Section 5.2 #8, 9, 15, 20, 27, 30; Section 5.3 #4, 6, 9, 13, 16, 28, 31; Section 5.4 #7, 18, 21, 23, 25, 29, 40

Infinite Geometric Series

Foundations of Arithmetic

Expected Value and Variance

Math Lecture 23

4.1. Probability Density Functions

Queueing Networks II Network Performance

THE KENNESAW STATE UNIVERSITY HIGH SCHOOL MATHEMATICS COMPETITION PART I MULTIPLE CHOICE NO CALCULATORS 90 MINUTES

Statistics and Probability Letters

Problem Set 9. Figure 1: Diagram. This picture is a rough sketch of the 4 parabolas that give us the area that we need to find. The equations are:

Solution of Tutorial 5 Drive dynamics & control

Definition of Tracking

SUMMER KNOWHOW STUDY AND LEARNING CENTRE

Math Calculus with Analytic Geometry II

Please review the following statement: I certify that I have not given unauthorized aid nor have I received aid in the completion of this exam.

Partially Observable Systems. 1 Partially Observable Markov Decision Process (POMDP) Formalism

Definite Integrals. The area under a curve can be approximated by adding up the areas of rectangles = 1 1 +

Structure and Drive Paul A. Jensen Copyright July 20, 2003

Week3, Chapter 4. Position and Displacement. Motion in Two Dimensions. Instantaneous Velocity. Average Velocity

Applied Stochastic Processes

DATA Search I 魏忠钰. 复旦大学大数据学院 School of Data Science, Fudan University. March 7 th, 2018

ACTM State Calculus Competition Saturday April 30, 2011

Fig. 1. Open-Loop and Closed-Loop Systems with Plant Variations

The strict priority scheduler

Problem Set 6: Trees Spring 2018

Proof that if Voting is Perfect in One Dimension, then the First. Eigenvector Extracted from the Double-Centered Transformed

Vectors , (0,0). 5. A vector is commonly denoted by putting an arrow above its symbol, as in the picture above. Here are some 3-dimensional vectors:

Pre-Calculus TMTA Test 2018

12.1 Introduction to Rational Expressions

Lecture 3: Probability Distributions

CISE 301: Numerical Methods Lecture 5, Topic 4 Least Squares, Curve Fitting

Properties of Integrals, Indefinite Integrals. Goals: Definition of the Definite Integral Integral Calculations using Antiderivatives

Decision Networks. CS 188: Artificial Intelligence. Decision Networks. Decision Networks. Decision Networks and Value of Information

TCOM 501: Networking Theory & Fundamentals. Lecture 7 February 25, 2003 Prof. Yannis A. Korilis

Chapter 6 Notes, Larson/Hostetler 3e

Linearly Similar Polynomials

Transcription:

M/G//GD/ / System! Pollcze-Khnchn (PK) Equton L q 2 2 λ σ s 2( + ρ ρ! Stedy-stte probbltes! π 0 ρ! Fndng L, q, ) 2

2 M/M/R/GD/K/K System! Drw the trnston dgrm! Derve the stedy-stte probbltes:! Fnd L,L q,, q R repr people K mchnes + + K R R R R K R K R 2,...,,,!! 0,,...,, 0 0 π ρ π ρ π

Exmple! The Gothm Townshp Polce Deprtment hs 5 ptrol crs. A ptrol cr bres down nd requres servce once every 30 dys. The polce deprtment hs two repr worers,ech of whom tes n verge of 3 dys to repr cr. Bredown tmes nd repr tmes re exponentl.! E(# polce crs n good condton)! E(down tme for cr needng reprs)! Frcton of tme prtculr repr worer s dle 3

Queues n Seres s,µ s 2,µ 2 s,µ Jcson s Theorem: If! Interrrvl tmes re exponentl, rte λ! Servce tme t ech server s exponentl! Ech stge hs nfnte cpcty wtng room Interrrvl tmes for ech stge exponentl, rte λ If λ < s µ, ech stge s M/M/s /GD/ / system. 4

Exmple 2! The lst two thngs tht re done to cr before ts mnufcture s complete re nstllng the engne nd puttng on the tres. An verge of 54 crs per hour rrve requrng these two tss. One worer s vlble to nstll the engne nd cn servce 60 crs per hour. After the engne s nstlled, the cr goes to the tre stton nd wts for ts tres to be ttched. Three worers serve t the tre stton. Ech wors on one cr t tme nd cn put tres on cr n n verge of 3 mnutes. Both nterrrvl tmes nd servce tmes re exponentl. Determne:! E(queue length t ech wor stton)! Totl E(tme cr spends wtng for servce) 5

Open Queung Networs s,µ s 2,µ 2 r r 2 p r 3 s 3,µ 3 s,µ! If λ < s µ for ll sttons, r! ech stge s M/M/s /GD/ / system.! L L + L 2 + + L! λ r + r 2 + + r, L/λ 6

Exmple 3! TO servers: An verge of 8 customers rrve from outsde t server, nd n verge of 7 customers per hour rrve from outsde t server 2. Interrrvl tmes re exponentl. Server cn serve t n exponentl rte of 20 customers per hour, nd server 2 cn serve t n exponentl rte of 30 customers per hour. After completng servce t server, hlf of the customers leve the system, nd hlf go to server 2. After completng servce t server 2, ¾ of the customers complete serve, nd ¼ return to server.! ht frcton of the tme s server dle?! Fnd E(# customers t ech server)! Fnd E(tme customer spends n the system)! ht f server 2 could serve only 20 customers per hour? 7

M/G/s/GD/s/ System! L,L q,, q re of lmted nterest! Queue never forms so L q q 0! s /µ! Frcton of rrvls turned wy λπ s! L L s λ( π s )/µ! Erlng s Loss Formul: π s depends only on λ /µ 8

Exmple 4! An verge of 20 mbulnce clls per hour re receved by Newport Hosptl. An mbulnce requres n verge of 20 mnutes to pc up ptent nd te the ptent to the hosptl. The mbulnce s then vlble to pc up nother ptent. How mny mbulnces should the hosptl hve to ensure tht there s t most % probblty of not beng ble to respond mmedtely to n mbulnce cll? Assume tht nterrrvl tmes re exponentlly dstrbuted. 9

Queung Dscplnes! FCFS Frst Come Frst Serve! SIRO Servce In Rndom Order! LCFS Lst Come Frst Serve! E( FCFS ) E( SIRO ) E( LCFS )! Vr( FCFS ) < Vr( SIRO ) < Vr( LCFS ) 0

Prorty Queues! q E(type customer s wtng tme)! E(type cust tme n system)! L q E(# type customers n lne)! L E(# type customers n system)

2 M /G //NPRP/ / System q q q n q L L S E λ µ λ λ +,, ) )( 2( ) ( Then 2 If < n ρ 0, 0, ρ µ λ ρ

Prorty Queue Exmple! A copyng fclty gves shorter obs prorty over long obs. Interrrvl tmes for ech type of ob re exponentl, nd n verge of 2 short obs nd 6 long obs rrve ech hour. Let type ob short ob nd type 2 ob long ob. Then we re gven tht E(S)2 mn, Vr(S)2, E(S2)4 mn, Vr(S2)2. Determne the verge length of tme ech type of ob spends n the copyng fclty. 3

Prorty Q E.g. 2! Gothm Townshp hs 5 polce crs. The polce dept receves two types of clls: emergency (type ) nd nonemergency (type 2) clls. Interrrvl tmes re exponentlly dstrbuted, wth n verge of 0 emergency nd 20 nonemergency clls ech hour. Ech cll type hs exponentl servce s, wth men of 8 mnutes (ssume tht, on the verge, 6 of the 8 mnutes s trvel tme from the polce stton to the cll nd bc to the stton). Emergency clls re gven prorty over nonemergency clls. On verge, how much tme wll elpse between the plcement of nonemergency cll nd the rrvl of polce cr? 4

Customer-Dep tng Costs! Assgn prorty such tht! c µ c 2 µ 2 c n µ n! Shortest Processng Tme (SPT) Dscplne:! Let c! µ µ 2 µ n! /µ /µ 2 /µ n 5

6 M /G /s/nprp/ / System q q q q L L s s P λ µ λ µ +,, ) )( ( ) ( Then If < n ρ ρ s 0, 0, ρ µ λ ρ

Prorty Queue E.g. 3! Consder computer system to whch two types of computer obs re submtted. The men tme to run ech type of ob s /µ. The nterrrvl tmes for ech type of ob re exponentl, wth n verge of λ type obs rrvng ech hour. Consder the followng three stutons:! Type obs hve prorty over type 2 obs, preempt! Type obs hve prorty over type 2 obs, no preempt! All obs re servced on FCFS bss.! Under whch system re type obs best off? orst off? ht bout type 2 obs? 7

M /G //PRP/ / System Preemptve resume vs. preemptve repet ( / µ )( ), 0 0, λ µ 8