Instructor: Dr. Ken Christensen Department: Computer Science and Engineering Office location: ENB 319 Office hours: Monday, Wednesday, and Friday: 4:30pm to 5:30pm. Call or email
to schedule an appointment. Phone: 974-4761 Email: christen@csee.usf.edu Homepage: http://www.csee.usf.edu/~christen
Teaching assistant: None
Required textbook: There is a required custom textbook and problem set
available from ProCopy (cost is about $50). There is also a required software
package called CSIM (student version cost is $65). How to obtain CSIM will be
discussed in class. There will be assigned readings to complement the lectures.
These readings will come from handouts and/or material available from the
Web.
Catalog course description: This special topics course is an introduction
to discrete-event simulation for performance modeling of communication and
computer systems. At the completion of this course, a student will be able to
model a system and predict its performance.
Course objectives:
As a result of successfully completing this course, students will:
Understand the basic principles of performance modeling.
Be able to select and use appropriate performance metrics when modeling a
system.
Understand the basics of queueing theory including Little's Law, the
M/M/1 queue, and the Erlang equations.
Know how to collect and characterize performance measurement data.
Know how to generate workload using probability distributions and using
a trace.
Understand the basic concepts of a discrete event simulation model including
model components, flowchart, and event list.
Learn how to design and implement process-oriented simulation models using
CSIM.
Know how to determine when a simulation program can be terminated using
confidence intervals
Know how to compare two designs using statistical hypothesis testing.
Understand the modeling and analysis process from a project perspective and
how to define experiments and present results.
Course topics:
This course will cover the following topics:
Week 1: Introduction to performance modeling
Week 2: Review of probability theory
Week 3: Generating random values and workload
Week 4: Introduction to queueing theory
Week 5: Organization of a discrete-event simulation model
Week 6: Wrap-up of first half of semester, review, and mid-term exam
Week 7: CSIM part #1
Week 8: CSIM part #2
Week 9: Model validation, verification, and output analysis
Week 10: Wrap-up of full semester, review, and final exam
Grading: Students will earn a grade based on assignments, project,
midterm exam, and a comprehensive final exam. The grade breakdown is:
Assignments: 10% (five assignments due on May 23rd, May 30th, June 6th,
June 13th, and June 27th at the beginning of class). Assignments are graded
for submission and attempt of all problems. Full solutions will be posted.
Project: 30% (due on July 19th by 5pm - early submission is very highly
recommended, late submissions will not be accepted)
Midterm exam: 30% (held in class on June 20th)
Comprehensive final exam: 30% (held in class on July 18th)
The grading scale is "no worse than" (note that there are no "+" or "-"
grades):
A = 90% through 100%
B = 80% through 89%
C = 70% through 79%
D = 60% through 69%
F = Less than 60%
Course policies:
Late work cannot be accepted except in cases of verifiable
emergencies.
A grade of Incomplete ("I") will not be given for this class. Please plan
accordingly.
We will be observing all university policies regarding religious holidays
and disability policies. Any student with a disability who needs special
accommodations must bring a current Memorandum of Accommodations from the
Office of Student Disability Services (this is the prerequisite for
receiving accommodations). Accommodated examinations through the Office of
Student Disability Services require two weeks notice.
You may not record audio or take video of the lectures.
Out of courtesy to other students please make sure that you turn off, or
place in silent mode, your cell phone.
Academic Integrity/Academic Dishonesty: I expect students to be honest
and not cheat on their assignments, project, and exams. Students may work
together on assignments and the project with one other person in the class.
All student pairs must then submit one copy of the completed assignment and/or
project with both names on it. Both students will earn the same grade. The
exams must be completed without giving or accepting assistance from other
students. Any source code copied from another source must be credited as such.
Open source software used must maintain all headers and other information as
required by the Open source license used. I expect you to know the University's
policies on student conduct, academic dishonesty, etc. Please see the
University's Undergraduate Catalog regarding these policies. Students found
cheating in any form will receive an FF grade for the course.
