Computer Science 890CE - Constraint-Based
Agents
Course Outline - Fall 2014
Instructor: |
Office: CW 308.13 |
|
||
|
|
|
|
|
|
|
|
|
|
Method of Evaluation: |
Report |
30% |
|
|
Presentation
20% Implementation
30% Oral Exam
20% |
|
|
||
|
|
|
|
Recommended Documentation
Topics
Report Format
·
The report should be submitted ONLY in PDF (preferred) or MS
Word (.doc).
·
The report should be formatted as follows
1.
20 to 30 pages. 8 1/2 x 11 inches portrait format, with a
single column.
2.
Text must be single-spaced
3.
Times New Roman regular 12 pts (or similar).
·
The following structure should be followed
1.
Title page
1.
Title (should be clear, descriptive and not too long)
2.
Your name
3.
Your affiliation
4.
Your e-mail address
5.
Abstract: should be clear, descriptive, self-explanatory and
not longer than 250 words.
6.
list of keywords (3 to 5)
2.
Body of text (divided in sections and subsections)
3.
Conclusion
4.
Acknowledgements (if any)
5.
References
Presentation
·
The slides should be submitted ONLY in PDF or PPT (MS Powerpoint) at least 3 days before the scheduled
presentation.
·
25-30min for the presentation followed by questions and the
oral exam.
Project Topics
1.
Constraint Solving
for Combinatorial Reverse Auctions (CRA) (Shil)
A.
Formally define
the CRA (with its variants) as a constraint optimization problem.
B.
Study the
different exact and approximation methods for solving the CRA problem. These
methods include the following.
a.
Branch and Bound
b.
Evolutionary
techniques (GAs, ACOs...etc)
c.
Stochastic Local
Search (SLS) algorithms
C.
Using a
parallel/distributed architecture, implement and experimentally compare the
above techniques on CRA problem instances.
2.
Variable and Value
ordering Heuristics for Constraint Processing (Ket
Wei)
A.
Study the effects
of variable and value ordering heuristics on constraint problems (both decision
and optimization problems).
B.
Study the
different variable and value ordering heuristics
a.
Static ordering
heuristics
b.
Dynamic ordering
heuristics
c.
Ordering
heuristics based on information gathering and learning
C.
Using a
parallel/distributed architecture, implement and experimentally compare the
above techniques on constraint problem instances.
Up:Teaching