Algorithmic game theory (NDMI098) - tutorials


Time of the tutorial: Monday 10:40am, in the room S321.

Instructor: Martin Balko. E-mail: balko (AT) kam.mff.cuni.cz

Lectures.


Information:
up
  • The tutorials take place irregularly according to the preliminary schedule shown below.
  • You will get the credit if you solve enough problems assigned during the semester. Each problem is evaluated according to its difficulty.
  • Usually you will have three weeks to solve the problems and write your solutions. I will give you hints for the most difficult problems after two weeks. After I give you a hint for a problem, the amount of points that you can get for solving this problem reduces to half of the original maximum. If you hand in a better solution after the hint has been given, I will add half of the points from your previous solution to the points from your new solution. It is not possible to get any points for solving the problems after a deadline has passed.
  • You need at least one quarter of all available points to get credit for the course.
  • Generally, the more points you get, the easier the exam will be for you.
  • On the day of a deadline, your solutions should be handed in before the lecture.
  • If you wish to see your score on this page, please choose a nickname and write it on the paper with your solutions (as well as your name), or send it by e-mail. Without the nickname I will not make your score public.
  • Current scores:
Nickname:  1.1   1.2   1.3   1.4   2.1a   2.1b   2.2   2.3   3.1   3.2   3.3    4.1     4.2a     4.2b     4.3    Total   Credit 
 Max Perfect  2 1 3 2 2 2 2 3 2 2 2 3 2 2 2 32 ---
 Bj  2 1 3 2 2 2 2 2 2 2 3 2 1 26 YES
 Lux  2 1 3 2 2 2 2 2 2 2 3 2 25 YES
 Ciaphas Cain  2 1 2 2 2 2 2 2 2 3 2 1.5 2 25.5 YES
 nickname1  2 1 2 2 2 2 11 YES
 Skrblík  2 1 2 1.5 6.5
  
  • Homework assignment 1: [PDF] - assigned 8.10.2018, deadline 29.10.2018.
  • Homework assignment 2: [PDF] - assigned 29.10.2018, deadline 26.11.2018.
  • Homework assignment 3: [PDF] - assigned 3.12.2018, deadline 17.12.2018.
  • Homework assignment 4: [PDF] - assigned 17.12.2018, deadline 7.1.2019.

Tutorials:
up
  • First tutorial (8.10.2018): Introduction, Nash equilibria. We showed solutions for exercises 1, 2a, and 5 and the first part of 4. List of exercises: [PDF].
  • Second tutorial (29.10.2018): Lemke--Howson algorithm. Deadline for Homework assignment 1. We showed solutions for Homework assignment 1 and for exercises 1 (example of complementary pivoting), 2, and 3. List of exercises: [PDF].
  • Third tutorial (12.11.2018): Epsilon-Nash and correlated equilibria. We showed solutions for exercises 1, 2, 3, 4b, and first part of 5b. We proved that, in zero sum games, the cumulative loss of no-regret algorithms is (up to epsilon) at most the value of the game. List of exercises: [PDF].
  • Fourth tutorial (26.11.2018): Regret minimization. Deadline for Homework assignment 2. We showed solutions for Homework assignment 2 and for exercises 1, 2, 3, 4, and 5. List of exercises: [PDF].
  • Fifth tutorial (17.12.2018): Revenue maximization. Deadline for Homework assignment 3. We showed solutions for Homework assignment 3 and for exercises 1, 2, 3, and 4. List of exercises: [PDF].
  • Sixth tutorial (7.1.2019): Bulow--Klemperer theorem and multi-parameter mechanism design. Deadline for Homework assignment 4. We showed solutions for exercises 1, 2, and 3. List of exercises: [PDF].

Valid XHTML 1.0 Transitional