Analysis of M/M/n/K Queue with Multiple Priorities

Similar documents
CHAPTER-5 PERFORMANCE ANALYSIS OF AN M/M/1/K QUEUE WITH PREEMPTIVE PRIORITY

THE ISRAELI QUEUE WITH INFINITE NUMBER OF GROUPS

Multi-Operation Multi-Machine Scheduling

MODELING THE RELIABILITY OF C4ISR SYSTEMS HARDWARE/SOFTWARE COMPONENTS USING AN IMPROVED MARKOV MODEL

Queues and Queueing Networks

Discrete-time Geo/Geo/1 Queue with Negative Customers and Working Breakdowns

1 Gambler s Ruin Problem

Combining Logistic Regression with Kriging for Mapping the Risk of Occurrence of Unexploded Ordnance (UXO)

ECE 6960: Adv. Random Processes & Applications Lecture Notes, Fall 2010

page 1 This question-paper you hand in, together with your solutions. ================================================================

Hotelling s Two- Sample T 2

AMS10 HW1 Grading Rubric

Solved Problems. (a) (b) (c) Figure P4.1 Simple Classification Problems First we draw a line between each set of dark and light data points.

Efficient Approximations for Call Admission Control Performance Evaluations in Multi-Service Networks

Signaled Queueing. Laura Brink, Robert Shorten, Jia Yuan Yu ABSTRACT. Categories and Subject Descriptors. General Terms. Keywords

Periodic scheduling 05/06/

Outline. Markov Chains and Markov Models. Outline. Markov Chains. Markov Chains Definitions Huizhen Yu

MATHEMATICAL MODELLING OF THE WIRELESS COMMUNICATION NETWORK

CHAPTER 5 STATISTICAL INFERENCE. 1.0 Hypothesis Testing. 2.0 Decision Errors. 3.0 How a Hypothesis is Tested. 4.0 Test for Goodness of Fit

Improved Capacity Bounds for the Binary Energy Harvesting Channel

Modeling and Estimation of Full-Chip Leakage Current Considering Within-Die Correlation

A Quadratic Cumulative Production Model for the Material Balance of Abnormally-Pressured Gas Reservoirs

Algorithms for Air Traffic Flow Management under Stochastic Environments

Approximating min-max k-clustering

BOUNDS FOR THE COUPLING TIME IN QUEUEING NETWORKS PERFECT SIMULATION

A PEAK FACTOR FOR PREDICTING NON-GAUSSIAN PEAK RESULTANT RESPONSE OF WIND-EXCITED TALL BUILDINGS

Developing A Deterioration Probabilistic Model for Rail Wear

Understanding and Using Availability

Discrete Signal Reconstruction by Sum of Absolute Values

Optimal Design of Truss Structures Using a Neutrosophic Number Optimization Model under an Indeterminate Environment

Universal Finite Memory Coding of Binary Sequences

Scaling Multiple Point Statistics for Non-Stationary Geostatistical Modeling

A MONOTONICITY RESULT FOR A G/GI/c QUEUE WITH BALKING OR RENEGING

Distributed Rule-Based Inference in the Presence of Redundant Information

Model checking, verification of CTL. One must verify or expel... doubts, and convert them into the certainty of YES [Thomas Carlyle]

Pretest (Optional) Use as an additional pacing tool to guide instruction. August 21

2-D Analysis for Iterative Learning Controller for Discrete-Time Systems With Variable Initial Conditions Yong FANG 1, and Tommy W. S.

4. Score normalization technical details We now discuss the technical details of the score normalization method.

Dynamic-Priority Scheduling. CSCE 990: Real-Time Systems. Steve Goddard. Dynamic-priority Scheduling

Estimation of Separable Representations in Psychophysical Experiments

On the Relationship Between Packet Size and Router Performance for Heavy-Tailed Traffic 1

A Comparison between Biased and Unbiased Estimators in Ordinary Least Squares Regression

New Schedulability Test Conditions for Non-preemptive Scheduling on Multiprocessor Platforms

Understanding and Using Availability

Linear diophantine equations for discrete tomography

Improving Patent Examination Efficiency and Quality: An Operations Research Analysis of the USPTO, Using Queuing Theory.

AN OPTIMAL CONTROL CHART FOR NON-NORMAL PROCESSES

Traffic Engineering in a Multipoint-to-Point Network

Morten Frydenberg Section for Biostatistics Version :Friday, 05 September 2014

Approximate Dynamic Programming for Dynamic Capacity Allocation with Multiple Priority Levels

A Study of Active Queue Management for Congestion Control

Estimation of component redundancy in optimal age maintenance

A Qualitative Event-based Approach to Multiple Fault Diagnosis in Continuous Systems using Structural Model Decomposition

PERFORMANCE BASED DESIGN SYSTEM FOR CONCRETE MIXTURE WITH MULTI-OPTIMIZING GENETIC ALGORITHM

Adaptive estimation with change detection for streaming data

Delay Performance of Threshold Policies for Dynamic Spectrum Access

Research of PMU Optimal Placement in Power Systems

QUEUING MODELS AND MARKOV PROCESSES

NUMERICAL AND THEORETICAL INVESTIGATIONS ON DETONATION- INERT CONFINEMENT INTERACTIONS

Anytime communication over the Gilbert-Eliot channel with noiseless feedback

On Code Design for Simultaneous Energy and Information Transfer

A Recursive Block Incomplete Factorization. Preconditioner for Adaptive Filtering Problem

A Model for Randomly Correlated Deposition

Some Unitary Space Time Codes From Sphere Packing Theory With Optimal Diversity Product of Code Size

Shadow Computing: An Energy-Aware Fault Tolerant Computing Model

Power Aware Wireless File Downloading: A Constrained Restless Bandit Approach

ECE 534 Information Theory - Midterm 2

CHAPTER-II Control Charts for Fraction Nonconforming using m-of-m Runs Rules

CMSC 425: Lecture 4 Geometry and Geometric Programming

Prof. Daniel A. Menascé Department of Computer Science George Mason University

y p 2 p 1 y p Flexible System p 2 y p2 p1 y p u, y 1 -u, y 2 Component Breakdown z 1 w 1 1 y p 1 P 1 P 2 w 2 z 2

Cryptanalysis of Pseudorandom Generators

Mathematical Efficiency Modeling of Static Power Converters

A General Damage Initiation and Evolution Model (DIEM) in LS-DYNA

The non-stochastic multi-armed bandit problem

Time Domain Calculation of Vortex Induced Vibration of Long-Span Bridges by Using a Reduced-order Modeling Technique

arxiv: v1 [physics.data-an] 26 Oct 2012

Proof Nets and Boolean Circuits

Chapter 7 Sampling and Sampling Distributions. Introduction. Selecting a Sample. Introduction. Sampling from a Finite Population

AI*IA 2003 Fusion of Multiple Pattern Classifiers PART III

One-way ANOVA Inference for one-way ANOVA

Positive decomposition of transfer functions with multiple poles

Research Note REGRESSION ANALYSIS IN MARKOV CHAIN * A. Y. ALAMUTI AND M. R. MESHKANI **

Uncorrelated Multilinear Principal Component Analysis for Unsupervised Multilinear Subspace Learning

Statistical Multiplexing Gain of Link Scheduling Algorithms in QoS Networks (Short Version)

Answers Investigation 2

A Network-Flow Based Cell Sizing Algorithm

1-way quantum finite automata: strengths, weaknesses and generalizations

Lecture 6. 2 Recurrence/transience, harmonic functions and martingales

On the Role of Finite Queues in Cooperative Cognitive Radio Networks with Energy Harvesting

Supplementary Materials for Robust Estimation of the False Discovery Rate

Lower Confidence Bound for Process-Yield Index S pk with Autocorrelated Process Data

System Reliability Estimation and Confidence Regions from Subsystem and Full System Tests

EXACTLY PERIODIC SUBSPACE DECOMPOSITION BASED APPROACH FOR IDENTIFYING TANDEM REPEATS IN DNA SEQUENCES

SUMS OF TWO SQUARES PAIR CORRELATION & DISTRIBUTION IN SHORT INTERVALS

Cryptography Assignment 3

LINEAR SYSTEMS WITH POLYNOMIAL UNCERTAINTY STRUCTURE: STABILITY MARGINS AND CONTROL

Information collection on a graph

DETC2003/DAC AN EFFICIENT ALGORITHM FOR CONSTRUCTING OPTIMAL DESIGN OF COMPUTER EXPERIMENTS

Published: 14 October 2013

Transcription:

Analysis of M/M/n/K Queue with Multile Priorities Coyright, Sanjay K. Bose For a P-riority system, class P of highest riority Indeendent, Poisson arrival rocesses for each class with i as average arrival rate for class i Service times for each class are indeendent of each other and of the arrival rocesses and are exonentially distributed with mean / i for class i Both Non-reemtive and Preemtive Priority Service discilines are considered Since the service times are exonentially distributed (i.e. memory less), the results for reemtive resume and reemtive non-resume will be identical Coyright, Sanjay K. Bose

Solution Aroach Define System State aroriately Draw the corresonding State Transition Diagram with the aroriate flows between the states Write and solve the balance equations to obtain the system state robabilities Note that we have given here the solution aroach that may be taken to solve a queueing roblem of this kind. This has been illustrated with simle examles. More comlex cases may be similarly formulated and solved with a corresonding increase in the solution comlexity Coyright, Sanjay K. Bose 3 M/M/-/- Queue with Preemtive Priority For a P-riority queue of this tye, define the system state as the following P-tule where (n, n,,n P ) n i = Number of jobs of riority class i in the queue i=,..,p Note that the server will always be engaged by a job of the highest riority class resent in the system, i.e. by a job of class j with service rate j if n j and n j+ =...=n P =. Coyright, Sanjay K. Bose 4

We illustrate the aroach first for a -riority M/M// queue,,,,,3,,,, 3, Coyright, Sanjay K. Bose 5 The corresonding balance equations for the -riority M/M// queue will be given by (,,, ( ( (......... + ) = + + +, + ) = + ) = + ) =,,,,,, These may be solved to obtain the desired state robabilities We illustrate next how this aroach may be generalized to aly to queues with finite caacity and/or multile servers, Coyright, Sanjay K. Bose 6 3

-Priority M/M//3 Queue (Preemtive Priority) (An examle of a queue with finite caacity),,,,,3,,, 3, New arrivals will be lost if they come when the system is in any of the circled states State Transition Diagram for the -Priority M/M//3 Queue with Preemtive Priority Coyright, Sanjay K. Bose 7 The corresonding balance equations are (,,,,,,,,3 3, ( ( ( ( ( + ) =,,,,, + + ) =,,,,, + + ) =,,, + + ) = + + ) = = = + + ) = = =,,,3,, 3, Normalization Condition,,3, 3,, =,,,, Coyright, Sanjay K. Bose 8 4

We can solve for the state robability distribution by solving any nine of the ten balance equation along with the equation for the normalization condition Job loss robability (or the blocking robability) =, +, + 3, +,3 Other desired robabilities may also be found from these state robabilities. Some examles are - P{server busy serving low riority job} =, +, + 3, P{one high riority job in the system} =, +, +, Coyright, Sanjay K. Bose 9 -Priority M/M//3 Queue (Preemtive Priority) (An examle of a queue with finite caacity and multile servers),,,3,,,,, 3, New arrivals will be lost if they come when the system is in any of the circled states Solve in the usual manner for the system state robabilities Coyright, Sanjay K. Bose 5

M/M/-/- Queue with Non-reemtive Priority We can roose two different methods of reresenting the system state for a M/M/c/K queue of this tye with P riority classes. Aroach I : If P < c, then this aroach gives a more comact reresentation using a P-tule than the more general Aroach II given next. State Reresentation (n,,n P, s,.,s P ) where n j = number of jobs of class j in system j=,,p s k = number of servers currently busy serving jobs of riority class k k=,..,p Coyright, Sanjay K. Bose Aroach II : This requires a (P+c)-tule of the following form State Reresentation n,,n P, s,.,s c ) where n j = number of jobs of class j in system j=,,p s k = riority class of the service currently on-going at server k k=,..,c Note that n +...+n P K for a finite caacity system We have used the reresentation of Aroach II in the examle described subsequently Coyright, Sanjay K. Bose 6

-Priority M/M//3 Queue (Non-reemtive Priority) (An examle of a single server queue with finite caacity),,,,,,,,,,,,,,,,,,,3,,, 3,, State Transition Diagram Coyright, Sanjay K. Bose 3 The balance equations for this queue are (,,,,,,,,,,,, ( ( ( ( + ) = ( ( + + + + + +,, + ) = + ) = + ) = + ) =,, + ) = + ) =,,,,,,,,,,,,,,,,,,,,,,,3, 3,,.and. Coyright, Sanjay K. Bose 4 7

,,,,,,,,,3, 3,, = = = = = =,,,,,,,,,,,,,,,, with the following normalization condition,,,,,,,,,,,,,3,,,,,,, 3,,,, = Coyright, Sanjay K. Bose 5 These equations may be solved on the usual way to obtain the individual state robabilities as er the definition of the system state These state robabilities may then be used to comute other erformance arameters and robabilities that may be of interest. For examle, the blocking robability of this system will be given by (,3, 3,,,,,,,,,, ) Other, similar robabilities and erformance measures may also be calculated Coyright, Sanjay K. Bose 6 8

Aroach may be extended in the usual fashion to analyze other similar systems as follows - More than two riority classes Other buffer caacity values or even queues with infinite buffer caacities Different caacity limits for the different riority classes Queues with more than one server Coyright, Sanjay K. Bose 7 9