The answers are in boldface.

  1. What is a common value auction?

  2. What is the dominant strategy for a buyer in the English (first-price sealed-bid) auction?

  3. What is the main advantage of the Dutch auction over the first-price sealed-bid auction?

  4. What is a possible way to cheat, as a buyer, on the Vickrey auction?

  5. Why is it that an item will often sell for more money on Ebay than it does on a regular retail outlet?

  6. When does the winner's curse appear?

  7. What is the best way to clear the following set of bids so as to be locally efficient and provide a uniform price? This is a double auction. Consider only integer clearing prices.
    Buyer-2 6
    Buyer-3 4
    Seller-2 1

  8. Combinatorial auctions are useful because

  9. You are asked to implement an algorithm that finds the optimal solution to the mailman problem.

    A set of mailman start at a central location and must deliver a set of letters. At the end each mailman goes home. They incur a cost proportional to the distance traveled and receive a reward for each letter delivered.
    You instantly realized that this problem can be solved using a combinatorial auction where each mailman bids:

  10. If the following bids were entered in a combinatorial auction (taken from the slide used in class, the items for sale are Spyke, Nightcrawler, Storm, ShadowCat, Rogue, Wolverine, Cyclops, Xavier, Jean Grey).
    Bid # Bid price Bid items
    1 $5 Wolverine
    2 $4 Cyclops, Jean Grey
    3 $5 Rogue, Wolverine
    4 $7 Cyclops, Storm
    5 $7 Nightcrawler, Storm
    6 $15 All items
    Which set of bids clears this auction while maximizing revenue?

  11. In Sandholm's algorithm for winner determination in combinatorial auctions we construct a search tree where each path from the root to a leaf: (pick best)

  12. A possible way to make Sandholm's algorithm for winner determination faster would be to (the choice you pick must be one that the algorithm does not currently implement).

  13. In negotiations, the conflict deal is:

  14. In the monotonic concession protocol, I terminate when

  15. What is the Zeuthen strategy?

  16. An appealing aspect of the Zeuthen strategy is that

    It does not work as is for combinatorial auction, but maybe there is a way to modify it so that it does?

  17. What is reflectional symmetry?

  18. Say you prefer Oscar the most, then Elmo, then Rosita, and finally Grouch. You are asked to participate in a Borda count auction to elect the favorite. What should your vote look like?

  19. What does Arrow's Impossibility Theorem tell us?

  20. Andrew, Bob, and Charles are trying to decide what color outfit to wear. They each have a Red, Blue, and Pink outfits. Andrew will only wear something different from everyone else. Bob will wear Red if Charles wears Blue, otherwise he wears Pink. Charles has decided to wear Pink. If they use the filtering algorithm to solve this problem, what color combination will Andrew, Bob, and Charles wear.

  21. Given that ¬ (a ∧ b ∧ c) and ¬ (b ∧ a ∧ f) and (a ∨ f), what can you derive from these facts using the hyper-resolution rule?

  22. In the hyper-resolution based consistency algorithm, what information is contained within a nogood?

  23. In asynchronous backtracking each agent maintains a local-view which contains

  24. In asynchronous weak-commitment search an agent will reset its own priority whenever

  25. What is the problem in distributed constraint optimization?

  26. The Adopt algorithm first builds a DFS tree. This DFS tree has the special characteristic that

  27. The Adopt algorithm can solve (pick best)

  28. In the Adopt algorithm each node maintains

  29. You are using LRTA* to solve the 8-puzzle and I tell you about five specific tile configurations and their respective distance to the goal (that is, how many moves it takes to get from each one to the goal configuration). How can you incorporate this information into your algorithm?

  30. Does LRTA* always converge to the optimal solution?

    If it always moves to the min neighbor it can get neglect to explore states that might be better.

  31. If you were to use LRTA* to solve the maze problem (that is, have one agent find its way out of a 2-dimensional maze), what would be good heuristic to use?

  32. If two players using Fictitious play converge to a solution, what can we say about that solution?

  33. Under replicator dynamics each agent

  34. An evolutionary stable strategy is one that

    "superinfluenced". I like making up words.

  35. What is the moving target function problem?

    Can we have a Markov reality in a quantum world?

  36. In the CLRI theory, the error() captures

  37. A 2-level agent

    We discussed layered architectures in the first part of the semester.

  38. The revelation principle in mechanism design states that

  39. Under the Groves mechanism each agent should receive a payment that is equal to some function h() plus

  40. What is the purpose of the Groves mechanism in mechanism design?

Copyright © 2001 José M. Vidal. All Rights Reserved.