| Dec 06, 2001 |
EECS 598-5: Mathematical Modeling and Simulation Techniques for Networking
|
Instructor: Mingyan Liu Communication networks have become increasingly complex systems. With the rapid expansion of the Internet, it is important that we are equipped with proper tools to analyze and gain insight on the performance, dynamics, technical and social implications of various mechanisms used in the Internet. In this course we will study two classes of such tools: mathematical modeling and simulation. Both of them are widely used in networking. On the one hand mathematical modeling, via abstraction, can be tractable, fast and intuitive, as well as facilitate optimization and sensitivity analysis. On the other hand simulation can be much more detailed and can handle large-scale systems. In this course we will review mathematical modeling techniques based on a range of principles and examine their effectiveness, particularly, the relationship between the simplicity of a model and its usefulness. The study will be highly application-oriented, in that there is always a very clear realistic subject to be modeled, be it a protocol, a channel or a policy. Under the second theme of the course we will review the techniques of computer simulation, which is widely used to study complex systems, and also often used to validate mathematical models. We will NOT show how to use a particular simulation tool (e.g.,NS-2, OPNET), but will show the basic statistical and discrete event concepts underlying most simulation tools. We will also discuss how to use simulation in a scientific way. This course will consist of both lectures (i.e., presentation by the instructor) and discussions (i.e., discussion on assigned papers in class). The subjects covered in this course are as follows (the sequence is subject to change):- Modeling of multiple access channels (e.g., channel errors, IEEE802.11) - Optimal routing and blocking probabilities - Performance modeling of TCP - Congestion control, rate control and utility maximizing - Simulation and the Monte Carlomethod - Internet traffic and self-similarity For each of these subjects there will be a list of selected papers as reading assignment. The final grade will be based on - the summary review on paper reading assignment; - participation in class discussions; - a term project/paper and presentation
For more information, please visit http://www.eecs.umich.edu/~mingyan/598W02/ . |
| |
|
