Monday, February 28, 2011

Project plan

As I mentioned in class, your "final exam" will be to complete either a project plan or a course module.  The details are now available on the course website. 

Basically, if you're taking 145 you'll do a 5-10 page plan for the project that you'll pursue in the spring term, and if not, you'll do a new course module (lecture plus a few HW problems).  If the course modules are good I may adopt them for next year!

Wednesday, February 23, 2011

Lecture 9

Today we finished up routing games and then started lecture 9 -- ad auctions.  We'll be doing ad auctions for the next 2-3 classes...so we'll be covering a lot about them, but still it's only the tip of the iceberg.

Today we covered a little bit of historical perspective about them and next class we'll get into the theory.  We'll start with the basics of auction theory and then move to the special case of ad auctions and ad exchanges (where we'll do a little in class experiment to give you a feel for the problem).

Note -- we will be having a make-up lecture on Friday 2/25 (this friday) at 1-2:30 in Ann 213.  Please be sure to attend since we'll be covering important material. 

Also, we will be having only 1 office hours this week (since the HW is not due until next friday).  Raga will be covering office hours Thursday 7-9.  Please stop by if you have questions about the ad auction competition or about the theory part of HW6.

Monday, February 21, 2011

Homework 6

Homework 6 is now out on the course website. Do get started on problem 4 immediately - even though the competition doesn't start till Wednesday, you need to start setting up your campaigns and thinking about your advertising strategies asap so that you have time to contact me (Raga) in case you run into difficulties, and sort things out. Good luck!

Thursday, February 17, 2011

Rankmaniac finale

We had the "final search" for rankmaniac 2011 yesterday and here are the results.  I hope you had fun and learned a lot...I was certainly impressed with a lot of your ideas!

Google
1) Giordan & Jamie
2) Mihail & Nihar ***
3) Zach & Ye
4) Fred & Scott
5) Sasha & Alejandro
6) Isaac & Bo
7) TAs
8) Wen-Hao
9) Doris, Dai Wei, Wenqi
10) Claudia & Ben
11) Emil
12) Andrew
13) George & Matt

Bing
1) Fred & Scott ***
2) Doris, Dai Wei, Wenqi
3) Wen-Hao
4) Zach & Ye
5) TAs
6) Emil
7) Isaac & Bo
8) Sasha & Alejandro
9) George & Matt
10) Giordon &  Jamie
11) Claudia & Ben
?? Andrew
?? Mihail & Nihar

*** means top result for "rankmaniac"

Monday, February 14, 2011

Lecture 8

Today we started the "network economics" section of the course in earnest.  Our "warmup" is the case of "load balancing games".  These games are motivated by the fact that a data center / server farm is simply too big to optimally assign jobs, so instead we could consider just using a greedy heuristic of sending jobs to the best server they can go.  Well, this algorithm can be viewed as a game where the players are the jobs...and we spent the lecture starting the analysis of this game. 

Remember -- Wednesday is the conclusion of "rankmania".  We'll have the ceremonial search at 12:30 in Ann 213.  So, feel free to bring your lunch...but make sure someone from your group is there so that we can easily identify which pages belong to which group and assign grades.

Also, remember that the project proposals are due wednesday at 12:30 too, so bring them along with you.

Friday, February 11, 2011

Rankmaniac reports

The rankmaniac reports from Homework 4 were due yesterday. I found two Rankmaniac reports yesterday outside Adam's office, and none today. I hope the rest of you have turned in your reports electronically to Bose (the Guru of Homework 4). If you have not submitted your reports yet, perhaps because you forgot about the report altogether, please do so immediately. Remember, the report alone is worth 50 points. If you don't submit your reports by tonight, you may not be able to claim any credit at all for a late submission.

Tuesday, February 8, 2011

Lecture 7

Yesterday's lecture was on "cascading behavior in networks", a.k.a. epidemics over networks, a.k.a., information cascades, a.k.a., diffusion over networks, ...  It goes by many names.

We looked at a few "toy" models that I hope gave you a flavor for what cascades are and how network structure (specifically clustering) affects their growth.  We really only touched the tip of the iceberg however, and so there are lots of related questions that would make great project ideas on this topic.  I hope to see a bunch of ideas here when you turn in your project proposals.

Remember: HWs are due on Friday this week, and so office hours are Wednesday and Thursday.  I'll be having mine on Wednesday, so please come early and get started on the HW (even though it's an easy week).  Otherwise I'll get lonely...
 

Tuesday, February 1, 2011

Homework 5

Homework 5 is now up on the course web page.  This HW is purely to get you up to speed with game theory.  If you've seen it before you may find it easy, but hopefully it will be fun. 

If the problems seem hard, or use terms that you don't know, be sure to come to Bose's "Intro to Game Theory" lecture this wednesday during the normal class time.  If you know all the terms in the homework and it's easy for you, there's no need to come on wednesday.

As always, post your questions about the homework as comments to this post.

Project ideas

In yesterday's lecture Michelle gave a great overview of a bunch of different project ideas and useful tools for you to consider when looking for your own project ideas.  I'm excited to see what you all come up with! 

If you're still looking for a group for the project proposals, post a comment here and I'm sure you'll find others looking.