Computational Democracy: Algorithms, Game Theory, and Elections Steven Wolfman 2011/10/27.

Slides:



Advertisements
Similar presentations
Voting Systems Explained
Advertisements

Which units are you most interested in covering? Unit A –Management Science Unit B – Growth Unit C – Shape and Form Unit D – Statistics.
6.896: Topics in Algorithmic Game Theory Lecture 18 Constantinos Daskalakis.
Nash’s Theorem Theorem (Nash, 1951): Every finite game (finite number of players, finite number of pure strategies) has at least one mixed-strategy Nash.
Voting Methods Continued
Presented by: Katherine Goulde
Voting and social choice Vincent Conitzer
Algorithmic Game Theory Uri Feige Robi Krauthgamer Moni Naor Lecture 9: Social Choice Lecturer: Moni Naor.
Math 1010 ‘Mathematical Thought and Practice’ An active learning approach to a liberal arts mathematics course.
The Voting Problem: A Lesson in Multiagent System Based on Jose Vidal’s book Fundamentals of Multiagent Systems Henry Hexmoor SIUC.
NOTE: To change the image on this slide, select the picture and delete it. Then click the Pictures icon in the placeholder to insert your own image. CHOOSING.
Mark Wang John Sturm Sanjeev Kulkarni Paul Cuff.  Basic Background – What is the problem?  Condorcet = IIA  Survey Data  Pairwise Boundaries = No.
Social Networks 101 P ROF. J ASON H ARTLINE AND P ROF. N ICOLE I MMORLICA.
How “impossible” is it to design a Voting rule? Angelina Vidali University of Athens.
IMPOSSIBILITY AND MANIPULABILITY Section 9.3 and Chapter 10.
CS 886: Electronic Market Design Social Choice (Preference Aggregation) September 20.
VOTING SYSTEMS Section 2.5.
Math for Liberal Studies.  We have studied the plurality and Condorcet methods so far  In this method, once again voters will be allowed to express.
Voting Theory.
MA 110: Finite Math Dr. Maria Byrne Instructional Laboratory 0345 Lecture 10/31/2008.
Math for Liberal Studies.  In most US elections, voters can only cast a single ballot for the candidate he or she likes the best  However, most voters.
MAT 105 Spring  There are many more methods for determining the winner of an election with more than two candidates  We will only discuss a few.
Social Choice: The Impossible Dream Michelle Blessing February 23, 2010 Michelle Blessing February 23, 2010.
Excursions in Modern Mathematics Sixth Edition
What is your favorite food?. Preference Schedule A Preference Schedule is a way to represent the order in which people like (prefer) certain items. The.
Presidential Election 2011 The Who? Where? What? & How?
Motivation: Condorcet Cycles Let people 1, 2 and 3 have to make a decision between options A, B, and C. Suppose they decide that majority voting is a good.
Snick  snack CPSC 121: Models of Computation 2009 Winter Term 1 Rewriting Predicate Logic Statements Steve Wolfman, based on notes by Patrice Belleville.
Arrow's Impossibility Theorem Kevin Feasel December 10, 2006
1 Introduction to Computability Theory Lecture15: Reductions Prof. Amos Israeli.
1 Introduction to Computability Theory Lecture12: Reductions Prof. Amos Israeli.
CPS Voting and social choice
Social choice theory = preference aggregation = voting assuming agents tell the truth about their preferences Tuomas Sandholm Professor Computer Science.
Snick  snack CPSC 121: Models of Computation 2010 Winter Term 2 Rewriting Predicate Logic Statements Steve Wolfman, based on notes by Patrice Belleville.
We vote, but do we elect whom we really want? Don Saari Institute for Mathematical Behavioral Sciences University of California, Irvine, CA
1 Manipulation of Voting Schemes: A General Result By Allan Gibbard Presented by Rishi Kant.
Social choice theory = preference aggregation = truthful voting Tuomas Sandholm Professor Computer Science Department Carnegie Mellon University.
How is this math? Mathematics is essentially the application of deductive reasoning to the study relations among patterns, structures, shapes, forms and.
Social choice (voting) Vincent Conitzer > > > >
Overview Aggregating preferences The Social Welfare function The Pareto Criterion The Compensation Principle.
MAT 105 Spring  We have studied the plurality and Condorcet methods so far  In this method, once again voters will be allowed to express their.
Arrow’s Impossibility Theorem
CPS Voting and social choice Vincent Conitzer
Snick  snack CPSC 121: Models of Computation 2012 Summer Term 2 Rewriting Predicate Logic Statements Steve Wolfman, based on notes by Patrice Belleville.
1 Elections and Manipulations: Ehud Friedgut, Gil Kalai, and Noam Nisan Hebrew University of Jerusalem and EF: U. of Toronto, GK: Yale University, NN:
Chapter 15 Section 1 - Slide 1 Copyright © 2009 Pearson Education, Inc. AND.
Math for Liberal Studies.  We have seen many methods, all of them flawed in some way  Which method should we use?  Maybe we shouldn’t use any of them,
Great Theoretical Ideas in Computer Science.
Let’s take a class vote. How many of you are registered to vote?
Discrete Mathematics, 1st Edition Kevin Ferland Chapter 6 Basic Counting 1.
Social choice theory = preference aggregation = voting assuming agents tell the truth about their preferences Tuomas Sandholm Professor Computer Science.
The mathematics of voting The paradoxes of democracy.
Chapter 9: Social Choice: The Impossible Dream Lesson Plan Voting and Social Choice Majority Rule and Condorcet’s Method Other Voting Systems for Three.
Excursions in Modern Mathematics, 7e: 1.Conclusion - 2Copyright © 2010 Pearson Education, Inc. 1 The Mathematics of Voting CONCLUSION Elections, Fairness,
Social Choice Lectures 14 and 15 John Hey. Lectures 14 and 15: Arrow’s Impossibility Theorem and other matters Plan of lecture: Aggregation of individual.
Impossibility and Other Alternative Voting Methods
1.
Chapter 9: Social Choice: The Impossible Dream Lesson Plan
Chapter 10: The Manipulability of Voting Systems Lesson Plan
Social choice theory = preference aggregation = voting assuming agents tell the truth about their preferences Tuomas Sandholm Professor Computer Science.
Applied Mechanism Design For Social Good
Impossibility and Other Alternative Voting Methods
Introduction If we assume
Chapter 9: Social Choice: The Impossible Dream Lesson Plan
Alex Tabarrok Arrow’s Theorem.
Voting systems Chi-Kwong Li.
Voting and social choice
Electoral Reform in BC …….but, why? (or why not?)
CPS Voting and social choice
Chapter 9: Social Choice: The Impossible Dream Lesson Plan
Presentation transcript:

Computational Democracy: Algorithms, Game Theory, and Elections Steven Wolfman 2011/10/27

A voting system is software.

2001 BC General Election

Some other oddities: MB Federal and 2000 US Presidential Vancouver Municipal

MANY Other Algorithms Approval (hand-raising in class, often), Range (IMDB), Party List Proportional (Germany), Cumulative (??), Block (Vancouver municipal), Condorcet (UBC AMS, “ranked pairs” flavour),... But.. what do we really want out of a voting system?

“Independence of Irrelevant Alternatives” (Intuition. Thanks to Sidney Morgenbesser.) We have apple and blueberry. I’ll take the apple. Oh, wait! We have cherry, too! In that case, I’ll take the blueberry

Formally: Independence of Irrelevant Alternatives If under one set of votes, A beats B, then... A still beats B under another set of votes with the same relative rankings of A and B.

General Definition: “Pareto (In)Efficient” If a change in the solution can make everyone better off, then the solution is “Pareto inefficient”. GFEDCBA Candidate (“option”)VoterKey: Question: Which of these is Pareto Inefficient?

Formal: Pareto Efficient For any two candidates A and B, if all voters prefer A to B, A must beat B. BA Candidate (“option”)VoterKey:

Formal: Dictator d is a dictator iff for any set of votes, the outcome precisely matches d’s vote. All hail 4!

Arrow’s Impossibility Theorem (Kenneth Arrow, 1951  Nobel Prize) PE IIA + Dictatorship Let’s prove it. We assume: no ties in votes or outcome. This assumption is unnecessary. We assume: finite #voters, at least three candidates, election is a function. All three are necessary.

Let’s play with votes. 1.Run the election. 2.Go back in time and change the votes. 3.Rerun the election. (AKA: explore the result of the election function on various inputs.) Scenario: move all B votes to the top PE and all B at top  B wins top

Scenario: all B at bottom 1.Run election 2.Change votes 3.Rerun election PE and all B at bottom  B “wins” bottom

Scenario: B at top or bottom Here, A > B and B > C. Generally, if B is in the middle, something beats B and some-thing loses to B. (Key results will always apply to the general case.) Can B end up in the middle like this? Let’s play!

Scenario: C just above A PE  C beats A IIA  A (still) beats B IIA  B (still) beats C CONTRADICTION Remember: A beats B B beats C B at top or bottom in all votes  B wins top or bottom

Scenario: find who moves B up Reminder: if every voter puts B at top or bottom, B wins top or bottom. PE  left side has B at bottom, right side has B at top

Scenario: find who moves B up Reminder: if every voter puts B at top or bottom, B wins top or bottom. At some point as we flip B up, it will move to the top.

Scenario: find who moves B up Reminder: if every voter puts B at top or bottom, B wins top or bottom. At some point as we flip B up, it will move to the top.

Scenario: find who moves B up Reminder: if every voter puts B at top or bottom, B wins top or bottom. At some point as we flip B up, it will move to the top.

Scenario: find who moves B up Reminder: if every voter puts B at top or bottom, B wins top or bottom. At some point as we flip B up, it will move to the top. Let’s focus on the voter who “controls” B. (3 in this case, but someone in general.) Starting with the right side.

Scenario: 3 moves A above B Why is A at the top? Compare to when we discovered 3...

Scenario: find who moves B up Reminder: if every voter puts B at top or bottom, B wins top or bottom. At some point as we flip B up, it will move to the top. Let’s focus on the voter who “controls” B. (3 in this case, but someone in general.) This time we want the left side.

Look at the A/B rankings here versus... our “3 puts A above B” scenario. IIA, left says A > B  A > B In fact, all but 3 can rearrange their votes, if they don’t move B...

Scenario: free up all but B and 3’s A As long as we keep voters’ A/B ordering the same, A > B, by IIA. B beats C. Must it? Compare to when we discovered 3...

Scenario: find who moves B up Reminder: if every voter puts B at top or bottom, B wins top or bottom. At some point as we flip B up, it will move to the top. Let’s focus on the voter who “controls” B. (3 in this case, but someone in general.) Back to the right side.

Scenario: free up all but B and 3’s A But, must B > C?Yes, by IIA, since B/C rankings match.

Scenario: free up the B’s Reminder: As long as we keep voters’ A/B ordering the same, A > B and B > C by IIA. Now, everyone moves their Bs around, and 3 moves its A around but keeps it above C. IIA  A > C, but the only fixed ranking is 3’s A > C!

Brief pause for formality: In general, A and C are arbitrary options other than B. So, 3 is a dictator with respect to all relative orderings except those involving B. All (but B) hail 3! Rest of proof is a repeat with different candidate for B; skipping...

Arrow’s Impossibility Theorem (Kenneth Arrow, 1951  Nobel Prize) PE IIA + Dictatorship

More algorithms/game theory rangevoting.org Ka-Ping Yee zesty.ca

Some Other Contexts of Use Usability and interface design Security and voter verifiability District creation and member allocation (and gerrymandering) Design specifications and standards Election auditing and fraud detection Voter files/databases Voter registration systems

Extra Slides (just in case)

Scenario: what about B Repeat proof, except with C at top/bottom rather than B. We find a dictator over B vs A (and B vs all but C), but remember...

Scenario: 3 moves A above B Because 3 moves A to the top, A > B. So, 3 is that dictator as well. And any candidate but B can “play” A. So, 3 is the dictator. All hail 3!

Canada’s Algorithm, Roughly 1.Start counts for each candidate and for “rejected” at 0. 2.For each ballot: a)If the ballot does not show a single candidate choice or has stray marks that might identify the voter, add one to the “rejected” total. b)Else, add one to the marked candidate’s total. 3.Report the results of all counts. After recounts, ties are do-overs: “deemed by-election” for the “deemed vacant” seat. Based on the Canada Elections Act. (Ties specified in the Parliament of Canada Act.)

A Harder Case: STV In Single-Transferable Voting, voters rank the candidates, a single election results in multiple candidates elected, and votes are transferred among candidates during the count based on preference order...