FA / EN

Computer Modeling and Performance Evaluation

This course will cover the fundamentals of the techniques of modeling and performance evaluation of computer systems. After familiarizing ourselves with the fundamentals like markov chains, birth death processes and queueing theory as well as the fundamentals of some machine learning techniques such as MDP, we will dig into applications that illustrate the use of these techniques. The applications covered will be from computer systems, computer networking and some topic in machine learning. We will also study computational techniques to complement analytical performance evaluation.

Course References
1 Computer Modeling and Simulation (In Persian)M. Vafaei Jahan, Islamic Azad University – Mashhad Branch Press, 2011.
2 Queueing Networks and Markov Chains, Modeling and Performance Evaluation with Computer Science Applications, G.Bolch, S.Greiner, Wily-Interscience, Second Edition, 2006.
3 Finite Markov Chain and Algorithmic Application, O. Haggstrom, Cambridge University Press, 2002.
4 Markov Processes for Stochastic Modeling, Oliver C. Ibe, Elsevier Press, 2009.
5 Markov Process and Applications, Algorithms, Networks, Genome and Finance, F.Pardoux, Wiley 2008.
6 Test Functions for Optimization Needs, M.Molga, C.Smutnicki, 2005.
7 Fuzzy Control of Queueing System, R. Zhang, Y. Phillis, V. Kouikoglou, 2010. 
Course Topics
Markov Process and Markov Chains ( 6 hours)
— Markov Chain Example I: Google Page Rank
— Ergodic Markov Chain (Film: Example with MATLAB, ErgodicMarkovChain)
— Non Ergodic Markov Chain (Film: MarkovNonErgodic01.mp4, NonErgodicMarkovChain)
Reversible Markov Chains ( 2 hours)
— Markov Chain Example 2: Markov Clustering
Birth and Death Process (4 hours)
Queue Theory (10 hours)
 – M/M/m/K Queue Model
 – Queue Network
#1 Select scientific paper for presentation (3 hours)
Hidden Markov Model (HMM) – (6 hours) –
HMM Example I
– HMM: Forward Algorithm (Farsi) with MATLAB
– HMM: Backward Algorithm (Farsi) With Matlab
HMM Baum Welch Example II
Markov Decision Process (MDP) – (6 hours)
– MDP Example
– MDP (Value Iteration Algorithm) – Video
#2 Paper Presentation
Instance of Exam and Practice
1 Load Testing of Web Server; Created by Dr. Majid Vafaei Jahan; 2012.
2 Final Exam 01; Spring Semester 1392-1393;
3 Final Exam 02: Fall Semester 1394-1395;
Reading List Papers
1 Load Testing of Web Sites, D.A. Menasce, D.A, Internet Computing, IEEE, Volume 6, Issue 4, pp:70 – 74, (2002) .
2 Evaluating Variable-Length Markov Chain Models for Analysis of User Web Navigation Sessions, Knowledge and Data Engineering, IEEE Transactions on,Volume: 19 Issue: 4, (2007).
3 Link Prediction and Path Analysis Using Markov Chains, R. Sarukkai, Computer Networks, Volume 33, Issues 1–6, pp:377–386, (2000).
4 An Efficient Algorithm to Rank Web Resources, D. Zhang, Y. Dong, Computer Networks, Volume 33, Issues 1–6, pp: 449–455, (2000)
5 A Web Log Mining System Based on MDP Algorithm, J.Gao, T.Yan and Y. Dang, Recent advances in Computer Science and Information Engineering, Volume 124, 447-453, (2012).
Course Evaluation
Each Exercise = +0.25
* Paper Presentation = 2 / 20
* Final Exam = 14 / 20
* Take Home and Represent as Paper Format = 3 / 20
Conference Paper = 2 / 20 (+1)