Comments on Lecture 4

08Nov07

I thought we went over some interesting results today…and our discussion turned out so provide some nice intuition about the contrast beween the behavior of a monopoly in the observable and the unobservable models.

If anyone missed class and would like to look over my notes before next time, just drop by my office.

A couple of things that came out of our discussion are:

1) what other extensions of Naor’s model have been studied?

2) What is the PoA in an M/G/1 unobservable queue, and why does it increase with job size variability?

3)  What extensions are of the unobservable model have been studied?

4) What is the PoA in other queueing-type congestions games (with more complicated structure than the parallel server case we discussed)?

Next up, Ho-Lin and Vera will teach us what is known along the lines of (4).

Filed under: Uncategorized   |  Leave a Comment