|
|
Michael Shalmon
Associate
Professor
Universite du Quebec, INRS-EMT
B.Sc (electrical engineering) (Technion-IIT),
Ph.D. (electrical engineering) (McGill University)
Tel: (514)
875-1266
shalmon at emt.inrs.ca
|
Unless otherwise noted everything is copyrighted and
scholarly posted for personal use only. Permission to reprint/republish for
advertising or promotional purposes or for creating new collective works for
resale or redistribution to servers or lists, or to reuse any copyrighted
component in other works must be obtained from the copyright holder.
Main
References
Bio
I
received the B.Sc degree in EE from the Technion-Israel Institute of Technology
in 1967 where I further studied communication theory in 1971-1972 while being
teaching assistant to Professors A. Lempel and J. Ziv.
From
1967 till 1980 I worked in industry, first in Israel and since 1975 in Canada.
In
September 1980 I returned to full time study at McGill University, from which I
received the M.Sc degree in 1981 and the Ph.D degree in February 1985, both in
EE. My Ph.D thesis, “Queues and packet multiplexing networks”, was done under
the supervision of Professor Michael Kaplan.
Since 1983 I have taught at INRS,
a research and graduate teaching institute affiliated with the University of
Quebec, and since June 1985 I have been Associate Professor there. INRS/EMT/Shalmon
During
2000-2001 I spent a sabbatical year at INRIA-Rocquencourt and at France-Telecom
R&D.
Teaching
Communication Networks (technology; teletraffic
modeling; analysis, estimation and control)
Performance Evaluation (stochastic models of
queueing; analysis, estimation and simulation)
Past:
Stochastic Estimation and
Simulation (fast generation, output analysis, variance reduction)
Stochastic Processes for Discrete Event Systems (Markov
chains, regenerative processes, queues)
Stochastic Signal Processing (Wold & spectral
representations, prediction & filtering)
Statistical Communication (elements of source and
channel coding)
Research
Overview
During
1986-1990 my main research topic was spatial statistical modeling and image processing
applications. Student work under my supervision included the following.
I returned since to my PhD research
topics: fundamental queueing analysis, stochastic modeling of flow processing packet multiplexing teletraffic
networks, and developing analytical tools for the evaluation and control of
performance (quality of service). Main References
- In a series of papers and communications starting with a paper
submitted in 1982, I put forward the regenerative Memoryless off – General
on * Variable rate (M-G*V) process as a stochastic model of packetized
flows and I obtained a comprehensive queueing analysis of a practically
relevant class of Packet
switching tandem & tree traffic concentrating networks. The queueing analysis includes queueing/delay
invariant network reductions for general tree networks concentrating arbitrary
traffic of constant size packets, and queueing equivalences to the M/G/1
queue for M-G*V packetized flows traffic.
- I discovered in 1983 and developed since, in a series of papers and
communications starting with my Ph.D thesis, the fundamental Ladders
/LCFS/Branching path decomposition of the queueing process, providing a clear and powerful queueing
analysis and having important applications in theoretical/applied probability
and in combinatorics. Especially for the M/G/1 (workhorse) queue, it provides a
detailed analysis of the busy-idle cycle, it is applicable to the queue with
work dependent service and/or arrival rate, and it has the power of analysis in
non-stationarity (finite horizon, random environment, etc), see Queueing applications. In
particular, I obtained analytical formulas determining the variance (=>
speed of convergence) of on-line estimators of the queueing performance and of
its gradient (sensitivity) with respect to arrival and/or service rate.
I am currently interested in queueing analysis in
non-stationarity, in processor sharing branching analysis and its connection to
LCFS, and in wireless telecommunication networks.
All material written by Michael Shalmon
on this site @ 2007-2010 Michael Shalmon. All rights reserved.
Main References
Overview Talks
[Talk-5] Michael Shalmon, “Sample path decompositions,
queueing analysis via LCFS-PR, and teletraffic multiplexing”, CanQueue’2003: http://www.fields.utoronto.ca/audio/03-04/canqueue/shalmon
[Talk-4] Michael Shalmon, “Queueing analysis via path decomposition applied to teletraffic multiplexing”, Hebrew U. of
Jerusalem, May 2003: 2003HebrewUJerusalem
[Talk-3] Michael Shalmon, “Sample path decompositions
and queueing analysis via LCFS-PR applied to teletraffic multiplexing”, Carnegie-Mellon
U., CS-OR Aladdin seminars, Oct 2002: http://www.aladdin.cs.cmu.edu/seminars/abstracts/shalmon.html
[Talk-2] M. Shalmon, “Random walk ladders, tree decompositions and
queueing analysis”, INRIA Sophia-Antipolis, June 1999: http://www-sop.inria.fr/mistral/seminaires/resumes/RES1999625.html
[Talk-1] Michael Shalmon, “Random walks, embedded branching processes and
queues”, McGill U. & Technion-IIT, May 1989.
M-G*V stochastic
modeling of packetized flows
Queueing/delay analysis of packet
switching tree & tandem traffic concentrating networks
[MuxNets-7] Michael Shalmon: “Queueing analysis for
priority multiplexing of aggregated VBR on-off regenerative traffic”, ITC20-2007,
submitted, M-GVpriorityqueueITC20.pdf.
[MuxNets-6] Michael Shalmon: “Queueing analysis of multiplexing
networks: arbitrary and aggregated VBR on-off regenerative traffic”, ITC20-2007,
submitted, M-GVpriorityqueuenetsITC20.pdf.
[MuxNets-5] Michael Shalmon: “Queueing analysis for polling and
prioritized service of aggregated regenerative variable rate On-Off traffic
sources”, Sigmetrics Performance Evaluation Review 29(3): 20 (2001), ACMLink-SigmetricsMAMA2001
[MuxNets-4] Michael Shalmon: “Queueing analysis for polling and
prioritized service of aggregated regenerative variable rate On-Off traffic
sources”: Conference & Workshop on Stochastic Networks 2000,
Wisconsin, StochNets2000
[MuxNets-3]
Michael Shalmon: “Traffic
multiplexing networks: queueing analysis”, IEEE-IT Information theory and
Networking Workshop, Metsovo, Greece, 1999, 1999IT-Netwkrs.pdf , IEEELink-MuxNets3
[MuxNets-2] Michael S. Shalmon, “Exact delay analysis of
packet switching concentrating networks”, IEEE Transactions on
Communications, 35, 1265-1271, 1987, 1987ConcentratingNetPollPrior.pdf , IEEELink-MuxNets2
[MuxNets-1] Michael Shalmon and
Michael A. Kaplan, “A Tandem
Network of Queues with Deterministic Service and Intermediate Arrivals”, Operations
Research, 32, 753-773, 1984, 1984ConcentratingNetFCFS.pdf , JSTORLink-MuxNets1
Fundamental queueing analysis via
Ladders/LCFS/Branching decomposition
[Ladd/LCFS/Branch-8] Michael
Shalmon, “Ladders, tree decompositions and queueing analysis via LCFS-PR”, IMS
Symp.Appl.Prob.,Banf, 2002: 2002IMSAppProb.pdf
[Ladd/LCFS/Branch-7] Michael
Shalmon, “Random walk ladders, tree decompositions, queueing analysis and
teletraffic multiplexing”, Bernoulli Stoch.Process.Appl. Conf., Cambridge,
UK, 2001. 2001BernoulliStchAppl
[Ladd/LCFS/Branch-6]
Michael Shalmon, “Random walk ladders, branching random trees, and the
temporal structure of queueing excursions”, IEEE-IT Information Theory and
Networking Workshop, StLouis, Missouri, 1995. 1995IT-Networks.pdf
[Ladd/LCFS/Branch-5]
Michael Shalmon, “Random walks, branching decompositions and queues”,
p222-223, Proc. ORSA-TIMS (INFORMS) Appl.Prob.Conf, Paris, 1993, 1993InformsApplProbParis
[Ladd/LCFS/Branch-4] Michael
Shalmon, “The moments of the regenerative estimators for the M/G/1 queue and
their interpretation as functionals of an associated branching process”, Bernoulli
Stoch.Process.Appl Conf., Haifa, 1991: 1991BernoulliStchAppl
[Ladd/LCFS/Branch-3] Michael Shalmon, “Random walks, embedded branching
processes and queues”, IMS Symp.Appl.Prob., Sheffield, U.K., 1989, 1989IMSconfSheff.pdf
[Ladd/LCFS/Branch-2b] Michael
Shalmon, “Level crossings, ladder variables, and stochastic decompositions in
queueing systems with Poisson arrivals”, Prob.Eng.Inform.Sciences, Jan
1989, submitted, 1989LCFS/MG1.pdf
[Ladd/LCFS/Branch-2a] Michael
Shalmon, “The M/G/1 queue length via the LCFS nonpreemptive discipline”, J.Appl.Prob., May 1988, submitted, 1988LCFS-NPR/MG1.pdf .
[Ladd/LCFS/Branch -1] Michael Shalmon, “The GI/GI/1 Queue and its variations via the LCFS
preemptive resume discipline”, Prob.Eng.Inform.Sciences, 2,
215-230, 1988. 1988LIFO/GIGI1&GM1&MG1.pdf
Queueing applications of the Ladders/LCFS/Branching
decomposition
Branching analysis of processor sharing
[ProcShar-1] Michael Shalmon, “Explicit formulas for the variance
of conditioned sojourn times in M/D/1-PS”,
Operations Research Letters, 35, 463-466, 2007, online Oct
2006, submitted 2004, prior 2003 submission to PEIS. VariancePS, ScienceDirectLink-ProcShar1
Arrival/service rate change, scaling, LR & IPA
analysis, queueing estimators variance
[QueueEstim-5] Michael Shalmon, “Analytical formulas for the variance of on-line
queueing estimations, for the batch size needed, and their application to
teletraffic multiplexing”, Symposium Performance Evaluation Computers
Telecommunications Systems 2003, 2003SPECTS.pdf SCSLink-QueEst5
[QueueEstim-4] Michael Shalmon, “How Fast is the Estimation of Performance and
of Performance Gradient for Teletraffic Multiplexing”, INFORMS-Telecom
Workshop, 2002, 2002INFORMSTelecom.pdf
[QueueEstim-3] Michael Shalmon, “How fast is the stochastic evaluation of
expected performance and of its parametric sensitivity”, Telecommunication
Systems, 2, 379-397, 1994, 1994Expect&IPAsensitivity.pdf , SpringerLink-QueEst3
[QueueEstim-2] Michael Shalmon, “Analytical Formulas for
the Variance of Queueing Estimators”, SIAM Workshop on Simulation, San
Francisco, 1993, 1993SiamWkshpSimul
[QueueEstim-1] Michael Shalmon and
Reuven Rubinstein, “Error analysis for regenerative queueing estimators
with special reference to gradient estimators via likelihood ratio”, Annals
of Operations Research, 36, 383-396, 1992, 1992WMoments&LRsensitivity.pdf , SpringerLink-QueEst1
Spatial statistical modeling and image
processing - Student work
J.F Richard: "Coding of hexagonally sampled images with
psychophysical considerations"; 1st prize award at Quebec's student
engineering competition, 1991.
P. Bergeot: "Wavelet coding of images”, 1990.
R. Dijkerman: "Generation and
restoration algorithms for Gaussian and binary Markov random fields: theory and
experiments", 1990.