Networks II
ARPITA GHOSH
Dept. Of Information Science,
Cornell University
Jan 22, 2015
Networks II
1
Welcome!
cfw_trumpets
Welcome to Networks II!
cfw_trumpets
Outline for today:
What: Course overview
How:
Course information
Class logistics
Credit compone

INFO 4220: Homework 2 Solution
1. (Binary preferences and Pareto-eciency.) First of all, be careful about which preference model
you assumed while solving the problemPareto eciency does not imply agents have strict rankorder preferences, although that is

INFO 4220: Homework 4
Due: Tuesday March 17, 2015 at 5pm
1. In markets with one-sided rank-order preferences, we know that there is always some agent who receives her top-ranked house in any Pareto-ecient
matching. (This follows from the result we stated

INFO 4220: Homework 1
Due: Friday, Feb 6 2015, at 5pm
1. (Conditional probabilities.) Charitable givinglike most other aspects
of societyhas gone online, with a number of online platforms such as
NetworkForGood that enable easy donations to a wide range o

INFO 4220: Homework 2
Due: Fri, Feb 20, 2015 by 5pm
1. (Binary preferences and Pareto-eciency.) Recall the binary preferences model from class. A matching
M is Pareto-ecient if it is not Pareto-dominated by any other matching, i.e., there is no other matc

INFO 4220: Homework 4 Solutions
1. (a) (5pts.) Suppose there is a man and a woman who are at the top of each others priority lists.
Does every stable matching contain the pair (m, w)?
Yes; heres why. Suppose to the contrary that we can nd a matching which

INFO 4220: Homework 1 Solutions
January 26, 2015
1. (Conditional probabilities.) This problem is a straightforward application
of Bayes rule. Let us use the following notation:
G: The charity is good,
B: The charity is bad,
+: a single donor decides to do

Markets with Initial Endowments: The Core
ARPITA GHOSH
Dept. Of Information Science,
Cornell University
Spring 2016
Markets with Initial Endowments: The Core
1
Announcements
HW3 will be posted by Friday 10am
Due next Friday March 11, 10am on CMS
Reminder:

Markets with Initial Endowments: The Core
ARPITA GHOSH
Dept. Of Information Science,
Cornell University
Spring 2016
Markets with Initial Endowments: The Core
1
Recap: One-sided markets with initial endowments
Markets with initial endowments: (ak , hk )k2c

Matching Markets with Non-transferable Utilities:
One-sided rank-order preferences
ARPITA GHOSH
Dept. Of Information Science,
Cornell University
Spring 2016
Matching Markets with Non-transferable Utilities: One-sided rank-order preferences
1
News you can

Matching Markets with Non-transferable Utilities:
One-sided preferences
ARPITA GHOSH
Dept. Of Information Science,
Cornell University
Spring 2016
Matching Markets with Non-transferable Utilities: One-sided preferences
1
Announcements
Late joiners (if any)

Matching Markets: One-sided Binary Preferences
ARPITA GHOSH
Dept. Of Information Science,
Cornell University
Spring 2016
Matching Markets: One-sided Binary Preferences
1
News you can use
Homework 1 will be posted on CMS by Friday morning
Due Friday Feb 12

Markets with Initial Endowments: The Core
ARPITA GHOSH
Dept. Of Information Science,
Cornell University
Spring 2016
Markets with Initial Endowments: The Core
1
Recap: One-sided markets with initial endowments
Markets with initial endowments: (ak , hk )k2c

Matching Markets with Non-transferable Utilities:
Rank-order preferences
ARPITA GHOSH
Dept. Of Information Science,
Cornell University
Spring 2016
Matching Markets with Non-transferable Utilities: Rank-order preferences
1
Recap: What were doing
Last week:

Markets with Initial Endowments: The Core
ARPITA GHOSH
Dept. Of Information Science,
Cornell University
Spring 2016
Markets with Initial Endowments: The Core
1
Recap: One-sided markets with initial endowments
Markets with initial endowments: (ak , hk )kcf

Markets with Initial Endowments: The Core
ARPITA GHOSH
Dept. Of Information Science,
Cornell University
Spring 2016
Markets with Initial Endowments: The Core
1
Useful News
Quick informal poll:
1
2
Pace:
[A] Very fast [B] Fast [C] Just right [D] Slow [E] Z

Emily Lu (el553) HW 2
Discussant: Megan Teeking & Arpit Sheth
1)
a) Bobs assignments are not always Pareto-efficient. By counterexample, suppose there
are two students, A and B, and two rooms, x and y. A finds both rooms acceptable, but B
only finds x acc

Emily Lu (el553), Discussed with Michelle Li and Megan Teeking
Networks II
February 11, 2016
HW 1
1)
a) A = bad provider; B = bad rating
P(B|A) = 0.95; P(B|A) = 0.01; P(A) = 0.005; P(A) = 0.995
P (BA )P( A)
0.95 0.005
P(A|B) =
=
'
'
0.95 0.005+0.995 0.01

INFO 4220: Homework 1 Solutions
February 13, 2016
1. (Conditional probabilities.) This problem is a straightforward application
of Bayes rule. Let us use the following notation:
G: The provider is good; B: The provider is bad,
+: a single positive rating

Emily Lu (el553), Discussed with Michelle Li and Megan Teeking
Networks II
February 11, 2016
HW 1
1)
a) A = bad provider; B = bad rating
P(B|A) = 0.95; P(B|A) = 0.01; P(A) = 0.005; P(A) = 0.995
!(#|%) !(%)
*.,- *.*
P(A|B) =
=
= 0.3231 for

Matching Markets with Non-transferable Utilities:
One-sided preferences
ARPITA GHOSH
Dept. Of Information Science,
Cornell University
Spring 2016
Matching Markets with Non-transferable Utilities: One-sided preferences
1
News you can use
Homeworks:
HW2: Du

Matching Markets with Non-transferable Utilities:
One-sided rank-order preferences
ARPITA GHOSH
Dept. Of Information Science,
Cornell University
Spring 2016
Matching Markets with Non-transferable Utilities: One-sided rank-order preferences
1
News you can

INFO 4220: Homework 2
Due: Fri, Feb 26, 2016 by 10am
General instructions:
1. Do make sure to follow the instructions on writing up (counter)examples (CMS; Piazza).
2. The sage advice from the Hitchhikers Guide continues to hold (and always will!):
(a) Do

Matching Markets with Non-transferable Utilities
ARPITA GHOSH
Dept. Of Information Science,
Cornell University
Spring 2016
Matching Markets with Non-transferable Utilities
1
News you can use
Clickers: Register at
http:/atcsupport.cit.cornell.edu/pollsrvc/

Emily Lu (el553) HW 2
Discussant: Megan Teeking & Arpit Sheth
1)
a) Bobs assignments are not always Pareto-efficient. By counterexample, suppose there
are two students, A and B, and two rooms, x and y. A finds both rooms acceptable, but B
only fin

Markets with Initial Endowments: The Core
ARPITA GHOSH
Dept. Of Information Science,
Cornell University
Spring 2016
Markets with Initial Endowments: The Core
1
Announcements
HW3 will be posted by Friday 10am
Due next Friday March 11, 10am on CMS
Reminder:

Markets with Initial Endowments: The Core
ARPITA GHOSH
Dept. Of Information Science,
Cornell University
Spring 2016
Markets with Initial Endowments: The Core
1
Useful News
Quick informal poll:
1
2
Pace:
[A] Very fast [B] Fast [C] Just right [D] Slow [E] Z

Markets with Initial Endowments: The Core
ARPITA GHOSH
Dept. Of Information Science,
Cornell University
Spring 2016
Markets with Initial Endowments: The Core
1
News you can use
Assignment: Blog post
Deadline: Next Friday March 18th, 10am
Guidelines for po

Emily Lu
HW 4
1.
a. Yes, every stable matching must contain the pair (m,w). Lets say there is a stable
matching where m and w are not paired with one another. This would mean that
m and w prefer their current partners less than each other because they are

Markets with Initial Endowments: The Core
ARPITA GHOSH
Dept. Of Information Science,
Cornell University
Spring 2016
Markets with Initial Endowments: The Core
1
Recap: One-sided markets with initial endowments
Markets with initial endowments: (ak , hk )k2c