CS378: Autonomous Multiagent Systems -- Spring 2004: Assignments Page

Assignments for Autonomous Multiagent Systems (cs378)

Week 1 (1/20,22)

Readings:
  • Is it an Agent, or just a Program?: A Taxonomy for Autonomous Agents.
    Stan Franklin and Art Graesser
    •

    Programming: (due 12:30pm on Thursday, 1/22)

  • If you don't have a CS UNIX account (e.g. because you're a non-CS major), sign up for a temporary account IMMEDIATELY! It will take several days for the account to be activated. Also, please be aware that the account will expire and completely disappear two weeks into the spring semester.
  • The programming assignment has 3 parts:
  • Get Familiar with the soccer server.
  • Watch a game.
  • Create a game.
  • To turn in your log file use the turnin program with grader "mazda" and assignment label "prog1". When the assignment is there, send us an email to that effect.

    • Week 2 (1/27,29): Autonomous agents

    Readings:
  • Textbook: sections 1, 1.1 (pages 1-7), 2-2.6 (pages 15-36)
  • The RoboCup Synthetic Agent Challenge 97.
    Hiroaki Kitano, Milind Tambe, Peter Stone, Manuela Veloso, Silvia Coradeschi, Eiichi Osawa, Hitoshi Matsubara, Itsuki Noda, and Minoru Asada.
    Fifteenth International Joint Conference on Artificial Intelligence (IJCAI-97).
    PDF version, HTML version
  • Soccer Server Manual (click on "downloads).
    Note that the manual is not perfectly up to date. No need to read it from cover to cover. Just become familiar with it.
    •

    Exercises: (due 12:01am on Tuesday, 1/27)

  • Exercise 1 from Chapter 2 (p. 46).
    One agent not discussed in class or in the readings is sufficient.
  • Send your response as ASCII text (not encoded in any way) to pstone@cs.utexas.edu and mazda@cs.utexas.edu with subject: "Week 2 exercises".
  • As indicated on the course overview page, your response should be well-thought-out, coherent, and concise. Quality of written expression will be a factor in the grading (use full sentences). Short, to-the-point answers are preferred. For full credit, your email should be sent by 12:01am on Tuesday, 1/27.

    • Programming: (due at 12:30pm on Thursday, 1/29)

  • The programming assignment has 3 parts:
  • Score a goal.
  • 1 on 1.
  • Passing.
  • To turn in your files, use the turnin program with grader "mazda" and assignment label "prog2". When the assignment is there, send us an email to that effect with a brief description of your approach.

    • Week 3 (2/3,5): Agent architectures

    Readings:
  • Intelligence without Representation.
     Rodney A. Brooks.
    Artificial Intelligence 47 (1991), 139-159.
    PDF version.
  • Pages 1-9 of Structured Control for Autonomous Robots.
    Reid Simmons.
    IEEE Transactions on Robotics and Automation, 10:1, pp. 34-43, February 1994.
  • The CMUnited-98 Champion Simulator Team.
     Peter Stone, Manuela Veloso, and Patrick Riley.
    in RoboCup-98: Robot Soccer World Cup II, M. Asada and H. Kitano (eds.), 1999. Springer Verlag, Berlin.
    •  Exercises: (due 12:01am on Tuesday, 2/3)
  • Identify one way in which TCA departs from Brooks' design principles for his creatures.
    **AND** (optional)
    Send a free-form response to the readings (see the syllabus).
  • Send your responses as ASCII text (not encoded in any way) to pstone@cs.utexas.edu and mazda@cs.utexas.edu with subject: "Week 3 exercises".
    •

    Programming: (due at 12:30pm on Thursday, 2/5)

  • The programming assignment is to use communication among agents to help an agent improve its performance at some task such as keeping track of where the ball is.
  • To turn in your files, use the turnin program with grader "mazda" and assignment label "prog3". When the assignment is there, send us an email to that effect, with a brief description of your task, your communication protocol, and an answer to the following question: Could an opponent agent disrupt your communication method? How?

    • Week 4 (2/10,12): Multiagent systems

    Readings:
  • MultiAgent Systems.
     Katia Sycara.
    AI Magazine, 1998.
  • The above is an overview of multiagent systems. Another overview (optional):
    Multiagent Systems: A Survey from a Machine Learning Perspective.
     Peter Stone and Manuela Veloso.
    Autonomous Robots, volume 8, number 3, July 2000.
    •
  • Designing and Understanding Adaptive Group Behavior. (citeseer link)
    Maja J Mataric.
    Adaptive Behavior 4:1, Dec 1995, 51-80.
    • Exercises: (due at 12:01am on Tuesday, 2/10)
  • Respond to the readings in some way (free-form response). If you're stuck, you can answer the question from last year: Think of an application that could be implemented as a multiagent system or a single agent. Briefly describe the 2 approaches you envision and list/discuss some of their relative merits.
  • Send your responses as ASCII text (not encoded in any way) to pstone@cs.utexas.edu and mazda@cs.utexas.edu with subject: "Week 4 exercises".
    •

    Programming: (due at 12:30pm on Thursday, 2/19)

  • The programming assignment is to get familiar with the United-2002 code base (built on CMUnited-99) and use it to create a simple team capable of playing a full soccer game.
  • To turn in your files, use the turnin program with grader "mazda" and assignment label "prog4". When the assignment is there, send us an email to that effect.

    • Week 5 (2/17,19): Agent communication and Teamwork

    Readings:
  • Agent Communication Languages: The Current Landscape.
     Yannis Labrou, Tim Finin, and Yun Peng.
    IEEE Inteligent Systems, March/April, 1999.
  • On Team Formation. (If the link doesn't work for you, it's also available from citeseer - top right corner)
     Cohen, P. R., Levesque, H. R., and Smith, I.
    in Hintikka, J. and Tuomela, R. (Eds.) Contemporary Action Theory. Synthese, 1997.
  • Supplemental (optional) readings are on the class resources page. At least look at the abstracts to see if you're interested in reading them.
    • Exercises: (due at 12:01am on Tuesday, 2/17)
  • Respond to the readings in some way (free-form response). One acceptable response is to answer one of the questions from last year:
  • Programming assignment 3 was to use communication to help robotic soccer agents improve at some task. Will KQML, KIF, or FIPA ACL be useful to you as you work on the assignment? Why or why not?
  • Choose a domain or example not discussed in the readings and briefly describe how it could be represented in terms of joint intentions.
  • Send your responses as ASCII text (not encoded in any way) to pstone@cs.utexas.edu and mazda@cs.utexas.edu with subject: "Week 5 exercises".
    •

    Programming: (due at 12:30pm on Thursday, 2/19)

  • See "Week 4" above.

    • Week 6 (2/24,26): RoboCup case studies

    Readings:
  • Read all the ABSTRACTS and then choose ANY TWO (2) of the following RoboCup case studies (to help you think about your proposal):

  • Reactive Deliberation: An Architecture for Real-time Intelligent Control in Dynamic Environments. (If the link doesn't work for you, pdf version is also available from citeseer - top right corner)
     Michael K. Sahota.
    Proceedings of the Twelfth National Conference on Artificial Intelligence, 1994.
    (From the group that invented robotic soccer: pre-RoboCup)
  • Evolving Team Darwin United.
     David Andre and Astro Teller.
    in Asada, M. (ed) Robocup-98: Robot Soccer World Cup II. Springer-Verlag, Berlin, 1999.
    (1998 evolutionary learning approach)
  • The CMUnited-99 Champion Simulator Team.
     Peter Stone, Patrick Riley, and Manuela Veloso.
    in M. Veloso, E. Pagello and H. Kitano (eds.) RoboCup-99: Robot Soccer World Cup III. Springer Verlag, Berlin, 2000.
    (1998, 1999 champion)
  • Learning Situation Dependent Success Rates of Actions in a RoboCup Scenario.
     Sebastian Buck, Martin Riedmiller.
    Pacific Rim International Coference on Artificial Intelligence, 2000.
    (2000 and 2001 runner-up)
  • Using Machine Learning Techniques in Complex Multi-Agent Domains.
     Martin Riedmiller, Artur Merke
    in I. Stamatescu, W. Menzel, M. Richter and U. Ratsch (eds.), Perspectives on Adaptivity and Learning (2002), LNCS, Springer
    (A more recent paper with more details.)
  • Towards an Optimal Scoring Policy for Simulated Soccer Agents.
     Jelle R. Kok, Remco de Boer, Nikos Vlassis, and F.C.A. Groen.
    In G. Kaminka, P.U. Lima, and R. Rojas, editors, RoboCup 2002: Robot Soccer World Cup VI, pages 292-299, Fukuoka, Japan, 2002. Springer-Verlag.
    (2001 3rd place, 2002 4th place)
  • An Empirical Study of Coaching..
     Patrick Riley, Manuela Veloso, and Gal Kaminka.
    In H. Asama, T. Arai, T. Fukuda, and T. Hasegawa, editors, Distributed Autonomous Robotic Systems 5,pp. 215-224, Springer-Verlag, 2002.
    (2001 coach competition champion)
  • Multi-robot decision making using coordination graphs
     Jelle R. Kok, Matthijs T. J. Spaan, and Nikos Vlassis.
    In A.T. de Almeida and U. Nunes, editors, Proceedings of the 11th International Conference on Advanced Robotics, ICAR'03, pages 1124-1129, Coimbra, Portugal, June 30-July 3 2003.
    (2003 champion)
    • Exercises: (due at 12:01am on Tuesday, 2/24)
  • Free-form response.
    **OR**
    For each of the two articles you chose to read, list the strengths of the described approach with respect to that in the other article. That is, what aspects of the complete task does it focus on that are ignored by the other approach, what unique techniques are used, or what does it do particularly well.
  • Send your responses as ASCII text (not encoded in any way) to pstone@cs.utexas.edu and mazda@cs.utexas.edu with subject: "Week 6 exercises".
    •

    Final Project Proposal: (due at 12:30pm on Thursday, 3/4)

  • See the final project page for full details.

    • Week 7 (3/2,4): Swarms and self-organization

    Readings:
  • "Go to the Ant": Engineering Principles from Natural Agent Systems.
     H. Van Dyke Parunak.
    Annals of Operations Research, 75:69-101, 1997.
  • Trail-Laying Robots for Robust Terrain Coverage.
     J. Svennebring and S. Koenig.
    In Proceedings of the International Conference on Robotics and Automation (ICRA), 2003.
    • Exercises: (due at 12:01am on Tuesday, 3/2)
  • Free-form response.
    **OR**
    Think of an application not discussed in the readings that you think could be better implemented with lots of simple agents rather than a small number of more cognitive agents. Describe and compare the merits of the two possible approaches.
  • Send your responses as ASCII text (not encoded in any way) to pstone@cs.utexas.edu and mazda@cs.utexas.edu with subject: "Week 7 exercises".
    •

    Final Project Proposal: (due at 12:30pm on Thursday, 3/4)

  • See the final project page for full details.

    • Week 8 (3/9,11): Applications

    Readings:
  • BDI-agents: from theory to practice.
     A. S. Rao and M. P. Georgeff.
    In Proceedings of the First International Conference on Multiagent Systems (ICMAS), 1995.
  • Electric Elves : Applying Agent Technology to Support Human Organizations.
     Chalupsky, H., Gil, Y., Knoblock, C. A., Lerman, K., Oh, J., Pynadath, D., Russ, T. A., and Tambe, M.
    In proceedings of the International Conference of Innovative Application of Artificial Intelligence (IAAI'01), 2001.
    • Exercises: (due at 12:01am on Tuesday, 3/9)
  • Free-form response.
    **OR**
    Identify an idea from one of the 2 applications in the readings that could be applied to one of the 2 other applications as well, and explain how it might improve (or at least change) the system.
  • Send your responses as ASCII text (not encoded in any way) to pstone@cs.utexas.edu and mazda@cs.utexas.edu with subject: "Week 8 exercises".
    •

    Survey: (due at 12:30pm on Thursday, 3/11)

  • Complete the midterm course evaluation survey.
    (this is instead of a programming assignment this week!)

    • Week 9 (3/23,25): Game Theory

    Readings:
  • Textbook: chapter 6
    **AND/OR**
    Roger McCain's introduction to game theory. [alternative link] .
  • Beginning to "Games with More than One Equilibrium"
  • "Cooperative Games"
  • "Sequential Games"

    • McCain motivates game theory from an economic perspective involving people as the actors as opposed to the textbook which motivates it from the AI agent perspective. But the theory is the same.

    • Exercises: (due at 12:01am on Tuesday, 3/23)
  • Free-form response
    **OR**
    Identify an application or situation not discussed in the reading that could be modeled as a matrix game. Specify the game matrix and identify all the dominant strategies and (pure strategy) Nash equilibria, if any.
  • Send your responses as ASCII text (not encoded in any way) to pstone@cs.utexas.edu and mazda@cs.utexas.edu with subject: "Week 9 exercises".
    •

    Final Project Progress Report: (due at 12:30pm on Thursday, 4/8)

  • See the final project page for full details.

    • Week 10 (3/30,4/1): Game theory II + RoboCup Rescue

    Readings:
  • RoboCup Rescue: Search and Rescue in Large-Scale Disasters as a Domain for Autonomous Agents Research.
    Hiroaki Kitano, Satoshi Tadokoro, Itsuki Noda, Hitoshi Matsubara, Tomoichi Takahashi, Atsushi Shinjou, and Susumu Shimada.
    Proc. of IEEE Conf. on System, Man and Cybernetics, Dec. 1999.
    • Exercises: (due at 12:01am on Tuesday, 3/30)
  • Free-form response
    **OR**
    Consider your final project focus. Could you have the same research focus in rescue? Explain the reasons.
  • Send your responses as ASCII text (not encoded in any way) to pstone@cs.utexas.edu and mazda@cs.utexas.edu with subject: "Week 10 exercises".
    •

    Final Project Progress Report: (due at 12:30pm on Thursday, 4/8)

  • See the final project page for full details.

    • Week 11 (4/6,8): Agent modeling

    Readings:
  • The UT Austin Villa 2003 Champion Simulator Coach: A Machine Learning Approach.
    Gregory Kuhlmann, Peter Stone, and Justin Lallinger.
    Under review.
  • Recursive agent modeling using limited rationality. (get cached copy from upper right corner)
    Jose M. Vidal and Edmund H. Durfee.
    In Proceedings of the First International Conference on Multi-Agent Systems (ICMAS-95), pages 376-383, Menlo Park, California, June 1995. AAAI Press.
  • Tracking dynamic team activity.
     Milind Tambe.
    National Conference on Artificial Intelligence(AAAI), 1996.
    • Exercises: (due at 12:01am on Tuesday, 4/6)
  • Free-form response
    **OR**
    Think of a domain not in the readings in which you could benefit from agent modeling. Briefly outline approaches with and without modeling and explain what benefits you would expect in the modeling case.
  • Send your responses as ASCII text (not encoded in any way) to pstone@cs.utexas.edu and mazda@cs.utexas.edu with subject: "Week 11 exercises".
    •

    Final Project Progress Report: (due at 12:30pm on Thursday, 4/8)

  • See the final project page for full details.

    • Week 12 (4/13,15): Distributed rational decision making

    Readings:
  • Distributed Rational Decision Making. Focus on Sections 5.1 - 5.5, but at least skim the rest.
    Tuomas Sandholm.
    In the textbook Multiagent Systems: A Modern Introduction to Distributed Artificial Intelligence, Weiss, G., ed., MIT Press. Pp. 201-258.
    • Exercises: (due at 12:01am on Tuesday, 4/13)
  • Free-form response.
  • Send your responses as ASCII text (not encoded in any way) to pstone@cs.utexas.edu and mazda@cs.utexas.edu with subject: "Week 12 exercises".
    •

    Final Project: (due at 12:30pm on Tuesday, 5/4)
    Final Project Report: (due at 12:30pm on Thursday, 5/6)

  • See the final project page for full details.

    • Week 13 (4/20,22): Auctions

    Readings:
  • Selling Spectrum Rights.
     John McMillan.
    Journal of Economic Perspectives, 8(3):145-162, 1994.
  • The 2001 Trading Agent Competition. (The "earlier version" is fine)
    Michael P. Wellman, Amy Greenwald, Peter Stone, and Peter R. Wurman.
    Fourteenth Innovative Applications of Artificial Intelligence Conference(IAAI-2002)
  • Optional (interesting, but not required)
  • The Timing of Bids in Internet Auctions: Market Design, Bidder Behavior, and Artificial Agents. Artificial Intelligence.
     Axel Ockenfels and Alvin E. Roth.
    AI Magazine, 23(3):79-88, Fall 2002.
    • Exercises: (due at 12:01am on Tuesday, 4/20)
  • Free-form response
    **OR**
    Suggest a use for agents in the FCC spectrum auction design described in the first reading.
    **OR**
    Suggest a change in TAC that would make the game more interesting or realistic in some way. In either case, briefly motivate the need for your suggestion and describe how your change addresses this need.
  • Send your responses as ASCII text (not encoded in any way) to pstone@cs.utexas.edu and mazda@cs.utexas.edu with subject: "Week 13 exercises".
    •

    Final Project: (due at 12:30pm on Tuesday, 5/4)
    Final Project Report: (due at 12:30pm on Thursday, 5/6)

  • See the final project page for full details.

    • Week 14 (4/27,29): Entertainment Agents

    Readings:
  • A Social Reinforcement Learning Agent.
     Charles Lee Isbell Jr., Christian R. Shelton, Michael Kearns, Satinder Singh, and Peter Stone.
    Fifth International Conference on Autonomous Agents, 2001.

  • Read all the ABSTRACTS and then choose ANY ONE (1) of the following papers. If you have time and are interested, by all means read more. Otherwise, file them away and read the ones you're interested in when you have a chance. These are all super-fun papers!

  • BoB: an Interactive Improvisational Companion.
     Belinda Thom.
    Fourth International Conference on Autonomous Agents, 2000.
  • Tears and Fears: Modeling emotions and emotional behaviors in synthetic agents.
     Jonathan Gratch and Stacy Marsella.
    Fifth International Conference on Autonomous Agents, 2001.
  • Creatures: Artificial Life Autonomous Software Agents for Home Entertainment.
     Stephen Grand, Dave Cliff, and Anil Malhotra.
    Millenium technical report 9601, 1996.
    • Exercises: (due at 12:01am on Tuesday, 5/4)
  • Choose a chatbot from the the AAAI chatbot page and interact with it for at least 10 minutes. What is it able to do? What kinds of things get it totally hosed? NOTE THE EXTRA EMAIL ADDRESS BELOW!
  • Send your responses as ASCII text (not encoded in any way) to ear@cs.utexas.edu and pstone@cs.utexas.edu and mazda@cs.utexas.edu with subject: "Week 14 exercises".
    •

    Final Project: (due at 12:30pm on Tuesday, 4/27)
    Final Project Report: (due at 12:30pm on Thursday, 5/6)

  • See the final project page for full details.

    • Week 15 (5/4,6): Multiagent Learning

    Readings:
  • Layered Learning.
     Peter Stone and Manuela Veloso.
    Eleventh European Conference on Machine Learning, 2000.

  • Optional READ AT LEAST THE ABSTRACTS!
    If you're on top of the final project and are interested, by all means read more.
  • Methods for Competitive Co-evolution: Finding Opponents Worth Beating.
     Christopher D. Rosin and Richard K. Belew.
    Proceedings of the Sixth International Conference on Genetic Algorithms, 1995.
  • Scaling Reinforcement Learning toward RoboCup Soccer.
    Peter Stone and Richard S. Sutton.
    In Proceedings of the Eighteenth International Conference on Machine Learning, pp. 537-544, Morgan Kaufmann, San Francisco, CA, 2001.
    • Exercises: (due at 12:01am on Tuesday, 5/4)
  • Free-form response
    **OR**
    Think of a domain not in the readings in which layered learning could be usefully applied.
  • Send your responses as ASCII text (not encoded in any way) to pstone@cs.utexas.edu and mazda@cs.utexas.edu with subject: "Week 15 exercises".
    •

    Final Project: (due at 12:30pm on Tuesday, 5/4)
    Final Project Report: (due at 12:30pm on Thursday, 5/6)

  • See the final project page for full details.

    • [Back to Department Homepage]

    Page maintained by Peter Stone
    Questions? Send me mail