Spring 2002 - CS 518T: Heuristic Search and Constraint
Processing
Instructor: Weixiong
Zhang
Time: TuTh 1:00-2:30pm
Location: Lopata 302
Course Information
Textbook:
-
Weixiong Zhang, State-Space
Search: Algorithms, Complexity, Extensions, and Applications, Springer,
1999, ISBN 0-387-98832-7
-
Richard Korf, Heuristic Search, selected chapters will be handed
out in the class
-
Rina Dechter, Constraint Processing, selected chapters will be handed
out in the class
Prerequisites:
CS 241 and CS 511A or consent of the instructor
Brief Description:
The course has two main parts and covers the main topics of heuristic search
and constraint processing. The first part focuses on single-agent
heuristic search problems and algorithms. The second part considers
constraint problems and constraint processing techniques. The course
will cover basic and advanced search techniques as well as their performance
analysis. It will also provide ample examples of real-world problems
and applications.
Purposes and Goals:
This is a hands-on course for graduate students who want to learn and use
search techniques to solve difficult combinatorial search problems.
The course is organized to handle large, complex real problems. The
students will have opportunity to develop algorithms and write programs
to solve some selected search problems or their own problems from research.
If you are a graduate student of mine, then you must work on a problem
different from those in your research. But you can choose one that
is related.
Grading:
There will be no homework or exam but one big project for each student,
which requires reading some research papers and developing algorithms and
programs and carrying out tests. The final grade will depend on a
project presentation and the quality of the project. For example,
if your final report is a publishable paper, you will definitely get an
A+. If you develop some new, novel ideas for a difficult problem
and demonstrate they are promising with some experimental results, then
you get an A or A-. If you just read some papers but with no results,
then you have a C or below.
I may ask you to give demonstration of the algorithms and/or systems
you develop on real data.
Schedule
Tues, Jan. 8: Examples of combinatorial search problems and state
space. Brute-force search
Thur, Jan 10: Best-first search. Project
discussions
Tues, Jan. 15: Linear-space search
Thur, Jan. 17: Memory-bounded search
Tues, Jan. 22: Real-time heuristic search
Thur, Jan. 24: Design of heuristic functions
Tues, Jan. 29: Phase transitions
Thur, Jan. 31: Approximation by exploiting phase transitions
Tues, Feb. 5: Local search and simulated annealing
Thur, Feb. 7: Simulated annealing and parallel simulated annealing
Tues, Feb. 12: Algorithm evaluation and algorithm engineering (a topic
on how you do your test and more)
Thur, Feb. 14: Constraints and constraint networks
Tues, Feb. 19: Consistency and constraint propagation
Thur, Feb. 21: Look-ahead search strategies
Student presentations:
Tues, Feb. 26: Sharlee Climer (paper:
I-10)
Thur, Feb. 28: Donald Jasper (paper:
I-9) and Andrew Gilpin
(paper: I-1)
Tues, Mar. 5: Spring break, no class
Thur, Mar. 7: Spring break, no class
Tues, Mar. 12: Justin Domke (paper:
I-3) and Ananda Rangan (paper:
II-3)
Thur, Mar. 14: Sharath Cholleti (paper:
II-1) and
Olcan Sercinoglu (paper I-2)
Tues, Mar. 19: Eric Demello (paper
I-11) and
Zhongsheng Guo
(paper II-18)
Thur, Mar. 21: Prashanth Pappu (paper
I-8),
Xiaotao Zhang (paper
II-13) and Peng Wang (paper
II-19)
Tues, Mar. 26: Kenneth Swanson and
Guandong
Wang (paper II-7)
Student project presentations:
Thur, Mar. 28: Zhao Xing (paper
II-14) and Donald
Jasper
Tues, Apr. 2: Sharlee Climer, Andrew Gilpin
and
Ananda Rangan
Thur, Apr. 4: Sharath Cholleti and
Eric
Demello
Tues, Apr. 9: Xiaotao Zhang and
Justin Domke
Thur, Apr. 11: Zhongsheng Guo and
Prashanth
Pappu
Tues, Apr. 16: Olcan Sercinoglu
and
Peng Wang
Thur, Apr. 18: Kenneth Swanson,
Guandong
Wang and Zhao Xing
Collaboration Policy
You must do an independent project. You are welcome to discuss
with anybody about your project, the problems you try to handle and ideas
of solving them. However, you must do your own, independent work
for developing algorithms, writing programs, conducting tests and comparisons,
writing the final report and preparing the presentation.
Joint projects are also welcome, as long as they have independent, roughly
separable parts each of which requires sufficient amount of work.
However, the students in a joint project must clearly state in their final
reports and presentations what their individual contributions are.
Office Hours
I have a semi open-door policy for this course. You are welcome to
stop by any time on Tuesday and Thursday whenever my door is open.
For other time, please send me email (zhang@cs).
Reading Material
Teaching notes on constraint processing:
Papers for possible projects (check this often as I continuously add new
items to this list.)
-
I: Single agent search and optimization:
-
(SMA* paper) Stuart Russell ``Efficient
Memory-Bounded Search Methods.'' In Proceedings of the Tenth European
Conference on Artificial Intelligence, Vienna: Wiley, 1992. (I also have
a paper that talked about SMA* on graphs.)
-
R.E. Korf and Weixiong Zhang, "Divided-and-conquer frontier
search applied to optimal sequence alignment", Proc. 17-th National
Conf. on Artificial Intelligence (AAAI-2000), Austin, Texas, July 30-August
3, 2000, pp.910-916.
-
Wheeler Ruml, Heuristic
Search in Bounded-depth Trees: Best-Leaf-First Search.
-
Wheeler Ruml, Incomplete
Tree Search using Adaptive Probing, IJCAI-01, pp. 235-241.
-
Wheeler Ruml, Using
Prior Knowledge with Adaptive Probing, Working Notes of the AAAI 2001
Fall Symposium on Using Uncertainty within Computation, pp. 116-120.
-
Toby Walsh, Depth-bounded
Discrepancy Search, Proceedings of IJCAI-97, 1997.
-
Pedro Meseguer and Toby Walsh, Interleaved
and Discrepancy Based Search, Proceedings of ECAI-98, 1998.
-
Boyan, J. A. and A. W. Moore. "Learning
Evaluation Functions for Global Optimization and Boolean Satisfiability."
AAAI-98 (Outstanding Paper Award).
-
Kaindl, H. and Kainz, G. (1997) "Bidirectional
Heuristic Search Reconsidered", Volume 7, J. AI Research, pages 283-317.
-
D. Applegate, R. Bixby, V. Chvatal, and W. Cook, "TSP
cuts which do not conform to the template paradigm", in Computational
Combinatorial Optimization, editors M. Junger and D. Naddef, pages
261-304, Springer, New York, N.Y., 2001
-
A. Plaat, et. al., "Exploiting
graph properties of game trees".
-
II: CSP, optimization and related:
-
Weixiong Zhang, Phase transitions
and backbones of 3-SAT and Maximum 3-SAT, CP-2001, 2001
-
John Slaney and Toby Walsh, Backbones
in Optimization and Approximation, Proceedings of IJCAI-2001, 2001.
-
Olivier Dubois and Gilles Dequen, A
backbone-search heuristic for efficient solving of hard 3-SAT formulae,
IJCAI'01
-
Roberto J. Bayardo Jr. and Robert Schrag. Using
CSP Look-back Techniques to Solve Real-world SAT Instances. AAAI-97.
-
Chu Min LI & Anbulagan, "Look-ahead
versus look-back for satisfiability problems", CP97.
-
Chu Min LI, "A
constrained-based approach to narrow search trees for satisfiability",
Information processing letters 71(1999) page 75-80.
-
Patrick Prosser, The
Dynamics of Dynamic Variable Ordering Heuristics, Proceedings CP-98
-
Xinguang Chen and Peter van Beek. Conflict-directed
backjumping revisited. Journal of Artificial Intelligence Research,
14:53-81, 2001.
-
Grzegorz Kondrak and Peter van Beek. A
theoretical evaluation of selected backtracking algorithms. Artificial
Intelligence, 89:365-387, 1997
-
Dechter R. and Frost, D., "Backjump-based
Backtracking for Constraint Satisfaction Problems", in forthcoming
Artificial Intelligence.
-
Frost, D., and Dechter, R., "Look-ahead
value ordering for constraint satisfaction problems." In "International
Joint Conference on Artificial Intelligence" (IJCAI-95), Montreal, Canada,
August 1995, pp. 572-578.
-
Dechter R., and Pearl, J., "Network-based heuristics for constraint-satisfaction
problems." In Artificial Intelligence, Vol. 34 (1), December 1987, pp.
1-38. (I have a copy. Ask me if you want to make a copy of
it.)
-
Russian Doll
Search for Solving Constraint Optimization Problems (1996) Gerard
Verfaillie, Michel Lemaître, Thomas Schiex, AAAI/IAAI, Vol. 1
-
Kask, K., and Dechter, R., "Graph-based
methods for improving GSAT." In proceedings of "National Conference
of Artificial Intelligence" (AAAI-96), Portland, Oregon, August 1996
-
Kask K., "New Search Heuristics
For Max-CSP" In the proceedings of the "Sixth International Conference
on Principles and Practice of Constraint Programming" (CP2000), September,
2000, pp. 255-269
-
Kostas Stergiou and Toby Walsh. Encodings
of Non-binary Constraint Satisfaction Problems, Proceedings of AAAI-99
-
Stuart A Grant and Barbara M Smith. Modelling
Exceptionally Hard Constraint Satisfaction Problems, CP-97, pages 182-195.
-
Steve Joy, et. al., "A
branch-and-cut algorithm for MAX-SAT and weighted MAX-SAT"
-
A. Frisch and T. Peugniez, "Solving non-boolean satisfiability
problems with stochastic local search".