מקווה שנהנתם ונתרמתם מתהליך העבודה על הפרוייקטים. במהלך היום כל קבוצה תקבל אימייל עם משוב על הפרוייקט. הציונים הסופיים של הקורס יתעדכנו במידע האישי בקרוב.
שנה טובה וגמר חתימה טובה,
כנרת
Recent Forum Posts
You need to define all the relevant basic definitions that are relevant to the main results.
I agree. There is no need to give all the details of each proof.
Hi,
Another question. How broad should the survey should be in the case that the article is using previous results?
To be more specific, in our case one article we survey use mechanisms, VGC, revelation principle, DISC ant etc.
Should we insert state and define all of those formally or we can just use them with cite (assume the reader is aware of them)?
Thanks,
Alon
in discussion Discussions / Spring 2017 Forum » survey - too long proof
Hi,
proofs can make the survey exceed the 10 pages limit. We have few options what to do:
1. just give short proof sketch and reference to the article
2. survey less - less algorithm and results
3. exceed the limit
I think option one is the best. what do you think?
This was answered in class. If there is still an ambiguity I will be happy to clarify.
I didn't fully understand what will the survey work include.
Is there a written explanation somewhere?
can't delete it :(
It looks like question 3 assume prior knowledge, but I don't recall that we cover this material in class.
I found this material in last year scribes, lectures 5 and 6.
agttau-2016DOTwikidotDOTcom/scribe-notes
you welcome
userDSIC
If you wrote it using latex or word you can email it to me. If not, please submit the hard copy.
in discussion Discussions / Spring 2017 Forum » Assignment submittion
Hey,
Can we submit a scan of our solution somehow or we must hand it physiclly to Schriber?
1. Each agent indeed wishes to minimize her cost.
2. You are correct that the set of valid strategies for player i contains exactly the paths of length l_i. But note that, as was written in the question, an agent can use an edge multiple times. Meaning that if for example, l_i = 3, and there is a cycle of length 2: s -> v -> s for some node v in the graph, a valid strategy for that player would be s -> v -> s -> v (which contains only 2 edges, but uses (s,v) twice).
I wanted to verify that I'm understanding the question correctly:
The strategies available to player i are only paths that have a length of exactly l_i from s (meaning the game is defined only on graphs for which there are such paths for each player), and each player is seeking to minimize his cost?
in discussion Discussions / Spring 2017 Forum » Question about homework 1, Q3.7
minimize the expected makespan
in discussion Discussions / Spring 2017 Forum » Question about homework 1, Q3.7
1. either
2. the strategy of each agent is independent of the other agents (their q's, their strategies)
3. right, but you can interpret a mixed strategy as "splitting" the bits.
in discussion Discussions / Spring 2017 Forum » Looking for HW partner
Hey, I am looking for a HW partner. If someone is in need of one please send me a message :)
No, sorry. Already found one.
in discussion Discussions / Spring 2017 Forum » Question about homework 1, Q3.7
Also, in 3.7, what does it mean for strategies to "together minimize the makespan"?
- When executed, they always achieve the optimal makespan?
- The expected makespan is optimal?
- The expected makespan is minimal among all strategies with properties (1) and (3) (equilibrium and not depending on the other loads)?
Hey Shlomi, are you still looking for a partner?