CSCE 475/875 Multiagent Systems
Class Syllabus
Fall 2013
Instructor
Name:
Prof. Leen-Kiat Soh
E-mail:
lksoh@cse.unl.edu Phone: (402) 472-6738
Office:
122E Avery Hall Classroom: 112 Avery Hall
Office Hours: 1:30
– 2:30 PM MW Class Time:
3:30 – 4:45 PM MW
Website:
http://www.cse.unl.edu/~lksoh/Classes/CSCE475_875_Fall13/
Teaching
Assistant
Name: Ertong Zhang
E-mail: ezhang@cse.unl.edu
Class
Objectives
This class will introduce you to the research topic of
multiagent systems (MAS), including what a MAS is, what agents are, and what
are the disadvantages and advantages of such a system in what types of
applications. We will present some
background issues in distributed problem solving (e.g., constraint
satisfaction) and planning (e.g., optimization) and build on the noncooperative game theory.
Then, equipped with the background knowledge, we will look into learning
in multiagent systems and agent communication.
Then we will cover formal agent topics such as social choice, mechanism
design, auctions, and coalitional game theory.
In terms of applications, we will also look into multiagent
negotiations, emergent behaviors (such as ants and swarms), and Robocup technologies.
Time permitting, we will also look into
research in real-time coalition formation.
The course materials in this class are based on the
textbook and journal/conference papers.
Required
Background
Prerequisites: CSCE 310 (Data Structures &
Algorithms required), or instructor permission.
Text Book
Shoham, Y. and K. Leyton-Brown
(2011). Multiagent
Systems: Algorithmic, Game-Theoretic, and Logical Foundations, Cambridge
University Press.
Grading
Final grades in this class will be assigned based on
the following scale:
A: 94% -
100%
A-: 90% - 93%
B+: 87% - 89%
B: 83% - 86%
B-: 80% -
82%
C+: 77% - 79%
C: 73% - 76%
C-: 70% -
72%
D+: 67% - 69%
D: 63% - 66%
D-: 60% -
62%
F: below
60%
A+ is awarded to a student whose work and
understanding of the class prove to be exceptional.
There will be about 5-10 collaborative topic summaries
(15%), several game days (group) (25%), one mid-term examination (20%), one
seminar presentation (group) (10%), and one final project (group) (30%). In general, students registered for CSCE875
will be given additional questions or activities for most of the
assignments.
Collaborative
Topic Summaries
Collaborative topic summaries are written summaries—in
Wiki form—of topics that we cover in class.
Specific requirements will be given before each summary assignment. The students will be divided into several
groups and each group will collaborate to prepare one final version of the
topic summary. Further details will be provided prior to the
first assignment.
Educational
Study. A formal study will be conducted to evaluate
the impact of collaborative writing and Wiki group formation techniques on
student learning. Your decision to
participate (or not participate) in this study will not impact your grades in
the class. You will be asked to sign a consent form for us to use the data
tracked using Wiki in this study.
Seminars
The seminar presentation is for the students to
present technical papers (or advanced chapters and sections from the textbook)
in the area of MAS, agents, and distributed AI.
A list of papers will be provided to the students. Each presentation will involve a Q&A
session paneled by the presenters and moderated by the instructor; and all
groups are required to participate in Q&A as well. One thing unique about our seminars is that
students will be graded for their participation in the Q&A sessions – as
panelist on the day of their presentation, and as scholars asking questions on
other presentation days. Details of the
grading criteria will be provided on the assignment handout later.
Final
Project
The final project will be for a multiagent simulation
that aims to address an important MAS problem such as how local decisions made
by a large group of agents (greater than 500) can lead to meaningful emergent
behaviors that satisfy global goals or the ad hoc team playing challenge
problem. The simulation software that we
will use is Repast. See the Resources
page on our course website for the software.
This assignment will be graded in 2 parts: programming (50%) and report
(50%). Each group member receives the
same score for his or her group. Details
of the grading criteria will be provided on the assignment handout later. One unique requirement is that the project
requires each group to conduct experiments and write a comprehensive discussion
of results based on the POJI[1]
style. Each group will be required to
submit a proposal, a progress report, and a final project report.
Game Days
There will be 3-4 special game days planned for the
semester, e.g., Auction Day, Voting Day, Learning Day, and Reputation Day. On each of these days, you will be required
to pit what you have learned in the class against your classmates in various
contests. You participation on those
days will be evaluated. Your written
reports for those days will be counted as participation as well. For this semester, each game day will likely
to place over two lectures due to the number of groups and the time we have for
each lecture. Each game day will be
graded in 2 parts: contest (50%) and report (50%). Each group member receives the same score for
his or her group. We will also have a
Game Days League – the winner of the league will get to sign the “Ball of
Fame”.
The contest part will be graded in proportion to how
your group ranks at the end of the game day.
For example, the winning team will be given N points; the second-ranked
team will be given N-2 points; the third-ranked team will be given N-4 points;
and so on. N will be determined later in
the semester depending on the number of groups we have in the class.
The report part will be graded roughly in the
following manner:
40%
Pre-game strategies
30%
During-game observations and changes in strategies
30%
Post-game strategies and lessons learned
Disabilities
Students with disabilities are encouraged to contact
Christy Horn for a confidential discussion of their individual needs for
academic accommodation. It is the policy of the University of Nebraska-Lincoln
to provide flexible and individualized accommodation to students with
documented disabilities that may affect their ability to fully participate in
course activities or to meet course requirements. To receive accommodation services, students
must be registered with the Services for Students with Disabilities (SSD)
office, 132 Canfield Administration, 472-3787 voice or TTY.
Academic
Misconduct
Violations of academic integrity will result in
automatic failure of the class and referral to the proper university
officials. The work a student submits in
a class is expected to be the student’s own work and must be work completed for
that particular class and assignment.
Students wishing to build on an old project or work on a similar topic
in two classes must discuss this with both professors. Academic dishonesty includes: handling in
another’s work or part of another’s work as your own, turning in one of your
old papers for a current class, or turning in the same or similar paper for two
different classes. Using notes or other
study aids or otherwise obtaining another’s answers for an examination also
represents a breach of academic integrity.
Sanctions are applied whether the violation was intentional or not.
Academic dishonesty of any kind will be dealt with in
a manner consistent with the CS&E Department's Policy on Academic Integrity
(http://cse.unl.edu/undergrads/academic_integrity.php). You are expected to
know and abide by this policy.
Those who share their code and those who copy other’s
code will be penalized in the same way; both parties will be considered to have
plagiarized.
To help avoid these problems, please start assignments
early and seek help when you need it.
PLAGIARISM
OF ANY KIND IN THIS COURSE WILL RESULT IN A GRADE OF F.
[1] This is the term coined by Professor Soh for Presentation, Observation, Justification, and Implication. We will discuss this in class.