Spring 2016 Forum (new posts)

הגשת הפרויקט: הגשת הפרויקט

Thu, 11 Aug 2016 10:07:28 +0000

הי, איך אנחנו נדרשים להגיש את הפרויקט? אימייל?

ex2 problem5: Re: ex2 problem5

Tue, 17 May 2016 19:30:28 +0000

why? it shouldn't be a problem to analyze the other case

ex2 problem 4: Re: ex2 problem 4

Tue, 17 May 2016 19:12:47 +0000

I approve, thanks Tzahi

Questions about ex 2: Re: Questions about ex 2

Tue, 17 May 2016 19:11:13 +0000

yes, thanks Itay

ex2 problem 4: Re: ex2 problem 4

Tue, 17 May 2016 18:32:10 +0000

I'm not Israela BUT:
1. Yes. (Notice he lies in the valuation not in the allocation)
2. It's a second price auction (winner = highest bidder, payment for winner = second highest bid).

Questions about ex 2: Re: Questions about ex 2

Tue, 17 May 2016 15:55:52 +0000

Israela - can you please confirm?

Questions about ex 2: Re: Questions about ex 2

Tue, 17 May 2016 07:35:56 +0000

1. yes
2. yes

ex2 problem5: ex2 problem5

Tue, 17 May 2016 05:53:55 +0000

Can we assume that n>k?
Thanks

ex2 problem 4: ex2 problem 4

Mon, 16 May 2016 22:27:43 +0000

1. In the example we are required to give, in the case the bidder is cheating (submit multiple bids) he can lie as much as he wants (ant give some allocation not what it is really worth for him). Right?
Also, the utility in this case is calculated as v-p, when v is the value that the (chosen) allocation really worth for him, right?
2. What is the definition of Vickrey auction?
Thanks

Questions about ex 2: Questions about ex 2

Mon, 16 May 2016 18:09:23 +0000

1. In question 1, can we assume the bidders have different private valuations?
2. In question 3, is "surplus" the same as social welfare?

questions about lecture 3: Re: questions about lecture 3

Mon, 16 May 2016 16:57:12 +0000

אני מצטערת אם לא הייתי ברורה.
העניין הוא שיש לך שאלות והערות רבות על כל הרצאה. אם הפסדת כמה הרצאות, כנראה שכדאי להשתמש בספר על מנת להשלים את החומר, בנוסף לסיכומים (או במקום). שימוש במקור נוסף יפתור את הבעיה של חוסר דיוקים ושגיאות בסיכומים.
אין לי יכולת לענות על כל כך הרבה שאלות על כל הרצאה במסגרת הפורום. הפורום נועד בעיקר להבהרות לגבי תרגילי בית (אני רק בודקת תרגילים, ולא מתרגלת).
אם יש שאלה ספציפית אני משתדלת לעזור, אבל בכמות הזאת זה מעבר לscope של העבודה שלי.

questions about lecture 3: Re: questions about lecture 3

Sun, 15 May 2016 15:27:26 +0000

מכיון שחלק מהשאלות היו האם יש טעות במקומות מסוימים, משמע אני מבין שתתכן אפשרות שישנה טעות בטקסט, כך שאיני מבין למה צריך להזהיר אותי מכך שהאפשרות ישנה (פעמיים).
בתגובה לשאלות על הרצאה 1 ענית לי שהשאלות לא היו מובנות. כך שניתן היה להבין, שפירוט וניסוח מחדש של השאלות ייענה בתשובות. אם מראש לא היתה לך כוונה לענות לי על השאלות, אולי היה כדאי לצין זאת, שלא אטרח ואוסיף עוד שאלות…
ולגופו של עניין, פספסתי חלק מן ההרצאות שלא מבחירה, ונראה לי שהדרך הכי יעילה להשלים את החומר היא דרך סיכומי ההרצאות. כך שאיני מבין למה אי אפשר לקבל מענה לשאלותי, מה גם שבחלק מהשאלות התשובה היא כמה מילים בלבד.
אשמח לתשובה לפחות לפוסט זה.

questions about lecture 3: Re: questions about lecture 3

Sun, 15 May 2016 13:36:33 +0000

As I previously said, the scribe notes might contain mistakes, and should be used carefully.
You can refer to other sources, as Nisan's book (there is a link to a online version at the website).

Definition of VCG: Re: Definition of VCG

Sun, 15 May 2016 13:32:52 +0000

see for example: https://he.wikipedia.org/wiki/%D7%9E%D7%9B%D7%A8%D7%96_Vickrey-Clarke-Groves
(or any other source)

Question 5 part b: Re: Question 5 part b

Sun, 15 May 2016 13:22:18 +0000

my mistake, it is enough to prove for Clarke pivot

questions about lecture 3: questions about lecture 3

Sun, 15 May 2016 12:26:50 +0000

1. Definition 1 page 1 - It says (at the third line of the definition) that the equality holds for each player. Isn't that redundant, (and therefore wrong)? It seems that the equality holds only for the deviating player. (And the fact that each player can be the deviating player was already mentioned in the second line).
2. Equations (1) and (2) in page 2 - shouldn't the sum on the last operand be over all edges which are in P_i\P_i^? And therefore, shouldn't the text in the paragraph between the equations be "the last element of the second operand is canceled" (instead of the current version that says "not canceled").
3. What is the calculation that leads to the fact that the example for CE on page 5 is indeed CE?
4. I want to make sure I understand why the CCE on page 5 is CCE: In half of the profiles (in the support) player 1 shares his edge with another player, and on the other half he doesn't. Meaning, his cost expectation is 1/2*2+1/2*1=1.5 .
If player 1 deviates to the fixed strategy 0 then his cost expectation would be 1/72[6*2+6*3+3*1+9*2+3*1+9*2+36*1]=1.5 .
Is this calculation correct?
5. Inequality (4) in page 6 - Shouldn't it be equality? I mean, isn't the cost of a profile defined as the sum of the costs of all players?
6. Cost sharing games - what is the meaning of v in profile p? (the last line of the first paragraph in page 8)
7. In the potential definition on page 8 - shouldn't the last sum in this equation go from i=1 to i=p_e?
8. Lemma 16 page 10 - the lemma bounds (lower and upper bound) the potential function. Is this bound holds for all potential functions of fair cost-sharing games? Or just the one defined in page 8? And in general, is there only one potential function for each game ? (in case it exists)
9. the last equation on page 10 - Isn't the demand that outer sum goes over all edges in p^i redundant? If we want to run over all edges that were chosen (at least by one player) the demand p_e>=1 is enough. Also, it's confusing in the middle part of the equation when i has a slightly different meaning.

Thanks a lot

some questions regarding lecture 1: Re: some questions regarding lecture 1

Sun, 15 May 2016 06:11:06 +0000

Hi,
Regarding point #4 - you are correct, the implication is derived later. I'll try to find some time to update the scribes.

Question 5 part b: Re: Question 5 part b

Sun, 15 May 2016 04:43:14 +0000

What about the following situation:
My vcg is like clarke pivot but instead h(v_-I)=max(…), I'll take h=0, and assume we have only 1 item
By definition is vcg, with payment (when i is the winner)
p_i=0, p_j=v_i (j!=i)
It is dsic as we would assume
But bidding again for even 0 is very helpful, because losing help me gain money

questions about lecture 2: questions about lecture 2

Sat, 14 May 2016 20:43:57 +0000

1. What is the meaning of s_i and r_j at the of page 1 and at the beginning of page 2?
2. How problems (2) and (2') (pages 2-3) are considered to be LP? In standard LP problems the value we are trying to minimize/maximize is the dot product of the variables vector with some constant vector.
3. Does the corollary that "It is possible to compute a NE in polynomial time for every Zero-Sum game" (cor' 8 page 4) refers to 2-players games only? or to any amount of players?

some questions regarding lecture 1: Re: some questions regarding lecture 1

Sat, 14 May 2016 19:59:17 +0000

2. example 1 page 8 - when I calculate p (the probability in which the rows player plays B in NE) I get the result p=3/5, in contrast to the result in the scribe : p=2/5. So I'm trying to understand if there is is a mistake in my calculation, and if there is - what is it?
The calculation was done as follows:
We assume we get NE when player 1 (rows player) plays B with probability p and plays O with probability 1-p.
We can calculate p using the fact that each pure strategy choice of player 2 (column player) leads to the same utility expectation for player 2, meaning E[u_2| player 2 plays B] = E[u_2| player 2 plays O]. (The utility expectation for player 2 when he plays B is equal to the utility expectation for player 2 when he plays O). So we have 2p+0(1-p) = 0p+3(1-p) leading to p=3/5.
Where is the mistake here?

3. The function in Brouwer fixed-point theorem is from domain C to range C, when C is a subset of R^t. (So any element in C can be viewed as a vector of size t). The domain and range used in the proof of the existence of NE are not a subset of some R^t. (Any element in that domain is a "vector of vectors"). So the implementation of Brouwer fixed-point theorem seems to be not very trivial.
Anyway, I was asking: can we interpret a profile as a vector of size t=|S_1|+..+|S_n| (and then the using of Brouwer theorem will be more trivial)?

4. in page 10, after the minmax theorem (lemma 15) it said "Immediate implication: with mixed strategies, the order of the turns does not matter - the players will take the same strategy sets."
How does it implies from that lemma? (The lemma above is an inequality (unlike the minmax theorem in lecture 2) and it seems that the only thing implies from it is that each player is better playing second).

Thank you