of Bath

CM30174: E-commerce & agents

Department of Computer Science


The questions given here are not an attempt to provide complete structured exam questions, based around a theme, with the typical format of some bookwork, followed by some synthesis, followed by some extrapolation. Instead they address issues that may normally be components of such questions and should be taken in that spirit. The presence or omission of any topic carries no implication either way for the content of the exam.

  1. Name and characterise the commonly agreed properties of an agent.
  2. Constrast the costs and benefits of algorithmic, declarative and utility-driven approaches to programming an agent.
  3. Outline the subsumption architecture and explain its strengths and weaknesses.
  4. Describe the key characteristics of Steel's Martian explorer subsumption architecture and explain how it worked.
  5. Explain how utility functions were used in the TileWorld.
  6. Outline the contract net protocol and sketch a potential application of it.
  7. Explain the purpose of agent communication languages and agent content languages, explaining why both are necessary and giving examples of each.
  8. Summarise the four main auction protocols: Dutch, English, (first-price) sealed bid and Vickrey's.
  9. Why is Vickrey's protocol effective and what are its advantages? In what circumstances is its use not advisable?
  10. Describe the components, and the interactions between them, of a reactive agent.
  11. Describe the BDI architecture (define the acronym BDI first) and how this is realised in the PRS system.
  12. How might a plan library be of use and be used by an agent?
  13. Sketch the prototypical structure of a negotiation protocol.
  14. Describe how argumentation protocols extend negotiation protocols.
  15. Sketch the structure of a deductive reasoning agent, illustrate with an example and discuss the limitations of the logic-based approach.
  16. What is Gabbay's separation theorem and explain in detail how it becomes the core of the Concurrent METATEM execution model.
  17. Summarise the key elements of temporal logic.
  18. In what situations might mobile agents be appropriate technology, what are the major technical issues created by them and how might they be resolved.
  19. How might a blackboard help in cooperative task solving? What are the problems it creates and what linguistic solutions have been proposed to help?
  20. Classify the different kinds of speech act, giving examples of each one.
  21. Elaborate on three objections to multi-agent systems.
  22. Explain the differences between deductive reasoning and reactive agents.
  23. Describe in a formal way action selection as theorem prover.
  24. Sketch Yoav Shoham's agent-oriented programming
  25. Give a formal description of planning.
  26. Describe the advantages and disadvantages of different types of layerings of hybrid agents.
  27. What is the relation between preference and utility?
  28. Can dominant strategies help finding Nash equilibria?
  29. What are zero-sum encounters?
  30. Give an example of a game without Nash equilibria.
  31. Some people to do not agree with the Nash equilibrium of the prisoner's dilemma. Give two of them and explain
  32. Explain English and Dutch auctions.
  33. Describe the Zeuthen strategy.
  34. Sketch logic-based argumentation.
  35. Why is modal logic an excellent logic for modelling multi-agent systems?

Julian Padget, Department of Computer Science, University of Bath