Piotr Faliszewski University of Rochester J ö rg Rothe Institute f ü r Informatik HeinrichHeineUniv D ü sseldorf Lane A Hemaspaandra University of Rochester Edith Hemaspaandra Rochester Institute of Technology ID: 229453
Download Presentation The PPT/PDF document "The Complexity of Llull’s Thirteenth-C..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Slide1
The Complexity of Llull’s Thirteenth-Century Election System
Piotr FaliszewskiUniversity of Rochester
Jörg RotheInstitute für InformatikHeinrich-Heine-Univ. Düsseldorf
Lane A. HemaspaandraUniversity of Rochester
Edith HemaspaandraRochester Institute of Technology
The Hebrew University of Jerusalem, June 17, 2009
Henning Schnoor
Rochester Institute of TechnologySlide2
Outline
IntroductionComputational study of electionsBribery and controlLlull/Copeland Elections
Model of electionsRepresentation of votes Llull/Copeland ruleResultsBribery and microbriberyControl of electionsManipulation
Hi, I am Ramon Llull. I have come up with the election that these guys now study!Slide3
Introduction
Computational study of electionsApplications in AIMultiagent systemsMulticriterion decision making
Meta search-enginesPlanningApplications in political scienceComputational barrier to cheating in electionsManipulationBribery
Control
Computational agents can systematically analyze an election to find optimal behaviorSlide4
Introduction
Many ways to affect the result of electionEvildoer wants to make someone a winner or prevent someone from winningEvildoer knows everybody else’s votes
ManipulationCoalition of agents changes their voteto obtain their desired effect
BriberyExternal agent, the briber, chooses a group of voters and tells them what votes to castThe briber is limited by budget
ControlOrganizers of the election modify theirstructure to obtain the desired result
In my times it was enough that we all promised we would not cheat...Slide5
Introduction
Manipulation versus bribery and controlControl attempted by the chair
Bribery attempted by some outside agentManipulation attempted by the voters themselvesSelf-interested voters
Bribed voters might choose to not follow what the briber said (unless the voting is public, or maybe not even then)
Okay… I can take the money, but I will vote as I like anyway!Slide6
Outline
IntroductionComputational study of electionsBribery and control
Llull/Copeland ElectionsModel of electionsRepresentation of votes Llull/Copeland ruleResultsBribery and microbriberyControl of electionsManipulation
Let me tell you a bit about my system....Slide7
Voting and Elections
Candidates and voters C = {c1, ..., c
n}V = {v1, ..., vm}
Each voter vi is represented via his or her preferences over C.Assumption: We know all the preferencesStrengthens negative results
Can be justified as wellVoting rule aggregates these preferences and outputs the set of winners.
Hi v
7
, I hope you are not one of those awful
people who support c
3
!
Hi, my name is v
7
.
How will they aggregate those votes?!Slide8
C = { , , }
Representing Preferences
Rational votersPreferences are strict total ordersNo cycles in single voter’s preference list
Example
> >Slide9
C = { , , }
Representing Preferences
Rational votersPreferences are strict total ordersNo cycles in single voter’s preference list
Not all voters are rational though!People often have cyclical preferences!Irrational voters are represented via preference tables.
Example
> >Slide10
C = { , , }
Representing Preferences
Rational votersPreferences are strict total ordersNo cycles in single voter’s preference list
Irrational preferences
Example
> >Slide11
C = { , , }
Representing Preferences
Rational votersPreferences are strict total ordersNo cycles in single voter’s preference list
Irrational preferences
Example
> >Slide12
C = { , , }
Representing Preferences
Rational votersPreferences are strict total ordersNo cycles in single voter’s preference list
Irrational preferences
Example
> >Slide13
C = { , , }
Representing Preferences
Rational votersPreferences are strict total ordersNo cycles in single voter’s preference list
Irrational preferences
Example
> >Slide14
C = { , , }
Representing Preferences
Rational votersPreferences are strict total ordersNo cycles in single voter’s preference list
Irrational preferences
> > >
Example
> >Slide15
Llull/Copeland Rule
The general ruleFor every pair of candidates, ci and cj,
perform a head-to-head plurality contest.The winner of the contest gets 1 pointThe loser gets zero points.At the end of the day, candidates with most points are the winners
Difference between various flavors of the Llull/Copeland rule?What happens if the head-to-head contest ends with a tie?Llull: Both get 1 point
Copeland0: Both get 0 pointsCopeland0.5
: Both get half a pointSlide16
Outline
IntroductionComputational study of electionsBribery and control
Llull/Copeland ElectionsModel of electionsRepresentation of votes Llull/Copeland ruleResultsBribery and microbribery
Control of electionsManipulation
Mr. Llull. Let us see just how resistant your system isSlide17
Bribery
E-bribery (E – an election system)
Given: A set of candidates C, a set of voters V specified via their preference lists, p in C, and budget kQuestion: Can we make p win via bribing at most k voters?E-$briberyAs above, but voters have prices and k is the spending limit.
E-weighted-bribery, E-weighted-$briberyAs the two above, but the voters have weights.
Hmm... I seem to have trouble with finding the right guys to bribe...Slide18
Bribery
E-bribery (E – an election system)
Given: A set of candidates C, a set of voters V specified via their preference lists, p in C, and budget kQuestion: Can we make p win via bribing at most k voters?E-$briberyAs above, but voters have prices and k is the spending limit.
E-weighted-bribery, E-weighted-$briberyAs the two above, but the voters have weights.
Result
Llull/Copeland rule is resistant (NP-hard; warning: a worst-case formalism) to all forms of bribery, both for irrational and rational voters
Mr. Agent. My system is resistant to bribery!Slide19
Microb
ribery
MicrobriberyWe pay for each small change we makeIf we want to make two flips on the preference table of the same voter then we pay 2 instead of 1Comes in the same flavors as briberyLimitationsCould be studied for
rational voters...... But we limit ourselves to the irrational
case.
We do not really need to change each vote completely...
Yeah... It’s easier to work with the preference Matrix™ ...
Preference table, I mean…Slide20
Microb
ribery
MicrobriberyWe pay for each small change we makeIf we want to make two flips on the preference table of the same voter then we pay 2 instead of 1Comes in the same flavors as briberyLimitationsCould be studied for
rational voters...... But we limit ourselves to the irrational
case.Results
Both Llull and Copeland0 are vulnerable to microbribery
Uh oh... How did they do that?!?!?Slide21
Microbribery in Copeland Elections
SettingC = {p=c0, c1,..., c
n}V = {v1, ..., vm}Voters vi are irrational
For each two candidates ci, cj:p
ij – number of flips that switch the head-on-head contest between them
ApproachIf possible, find a bribery that gives p at least B
points, ...
... and
everyone else at most B
points
Try all reasonable B’s
Validate B via min-cost flow problemSlide22
Proof Technique: Flow Networks
Notation:s(ci) – c
i score before briberyB – the point boundK – large number
capacity/cost
c
1
c
n
c
2
p
s
tSlide23
Proof Technique: Flow Networks
Notation:s(ci) – c
i score before briberyB – the point boundK – large number
capacity/cost
c
1
c
n
c
2
p
s
t
source
sink
mesh
source – models pre-
bribery scores
mesh – models bribery cost
sink – models bribery successSlide24
Proof Technique: Flow Networks
Notation:s(ci) – c
i score before briberyB – the point boundK – large number
capacity/cost
c
1
c
n
c
2
p
s
t
s(c
1
)/0
s(c
2
)/0
s(c
n
)/0
s(p)/0
B/0
B/K
B/K
B/K
source
sink
mesh
source – models pre-
bribery scores
mesh – models bribery cost
sink – models bribery successSlide25
Proof Technique: Flow Networks
Notation:s(ci) – c
i score before briberyB – the point boundK – large number
capacity/cost
c
1
c
n
c
2
p
s
t
s(c
1
)/0
s(c
2
)/0
s(c
n
)/0
s(p)/0
B/0
B/K
B/K
B/K
source
sink
mesh
1/p
10
1/p
21
1/p
20
1/p
2n
source – models pre-
bribery scores
mesh – models bribery cost
sink – models bribery successSlide26
Proof Technique: Flow Networks
Notation:s(ci) – c
i score before briberyB – the point boundK – large number
capacity/cost
c
1
c
n
c
2
p
s
t
s(c
1
)/0
s(c
2
)/0
s(c
n
)/0
s(p)/0
B/0
B/K
B/K
B/K
source
sink
mesh
1/p
10
1/p
21
1/p
20
1/p
2n
source – models pre-
bribery scores
mesh – models bribery cost
sink – models bribery successSlide27
Proof Technique: Flow Networks
Notation:s(ci) – c
i score before briberyB – the point boundK – large number
capacity/cost
c
1
c
n
c
2
p
s
t
s(c
1
)/0
s(c
2
)/0
s(c
n
)/0
s(p)/0
B/0
B/K
B/K
B/K
source
sink
mesh
1/p
10
1/p
21
1/p
20
1/p
2n
source – models pre-
bribery scores
mesh – models bribery cost
sink – models bribery successSlide28
Proof Technique: Flow Networks
Notation:s(ci) – c
i score before briberyB – the point boundK – large number
capacity/cost
c
1
c
n
c
2
p
s
t
s(c
1
)/0
s(c
2
)/0
s(c
n
)/0
s(p)/0
B/0
B/K
B/K
B/K
source
sink
mesh
1/p
10
1/p
21
1/p
20
1/p
2n
source – models pre-
bribery scores
mesh – models bribery cost
sink – models bribery successSlide29
Proof Technique: Flow Networks
Notation:s(ci) – c
i score before briberyB – the point boundK – large number
capacity/cost
c
1
c
n
c
2
p
s
t
s(c
1
)/0
s(c
2
)/0
s(c
n
)/0
s(p)/0
B/0
B/K
B/K
B/K
source
sink
mesh
1/p
10
1/p
21
1/p
20
1/p
2n
source – models pre-
bribery scores
mesh – models bribery cost
sink – models bribery successSlide30
Proof Technique: Flow Networks
Notation:s(ci) – c
i score before briberyB – the point boundK – large number
capacity/cost
c
1
c
n
c
2
p
s
t
s(c
1
)/0
s(c
2
)/0
s(c
n
)/0
s(p)/0
B/0
B/K
B/K
B/K
source
sink
mesh
1/p
10
1/p
21
1/p
20
1/p
2n
source – models pre-
bribery scores
mesh – models bribery cost
sink – models bribery success
Cost = K(n
(n+1
)/2
- p-score) + cost-of-bribery Slide31
Microbribery: Application
Round-robin tournamentEveryone plays with everyone elseBribery in round-robin tournaments
For every game there we knowExpected (default) resultThe price for changing itWe want a minimal price for our guy having most pointsRound-robin tournament exampleFIFA World Cup, group stage3 points for win
1 point for tie0 points for lossMicrobribery?Slide32
Microbribery: Application
Round-robin tournamentEveryone plays with everyone elseBribery in round-robin tournaments
For every game there we knowExpected (default) resultThe price for changing itWe want a minimal price for our guy having most pointsRound-robin tournament exampleFIFA World Cup, group stage3 points for win
1 point for tie0 points for lossMicrobribery?Applies directly!!
Given the table of expected results and prices …… simply run the Microbribery algorithmFor FIFA:
Simply use 1/3 as the tie value.Slide33
Outline
IntroductionComputational study of electionsBribery and control
Llull/Copeland ElectionsModel of electionsRepresentation of votes Llull/Copeland ruleResultsBribery and microbribery
Control of electionsManipulation
How will your system deal with my attempts to control, Mr. Llull...?Slide34
Control
Control of electionsThe chair of the election attempts to influence the result via modifying the structure of the electionConstructive control (CC) or destructive control (DC)
Candidate controlAdding candidates (AC)Deleting candidates (DC)Partition of candidates (PC/RPC)with or without the runoff
Voter controlAdding voters (AV)Deleting voters (DV)Partition of voters (PV)
My system is resistant to all types of constructive control!!
Okay, almost all. Slide35
Flavors of control
AC
– adding candidatesDC – deleting candidates(R)PC – (runoff) partition of candidatesAV
– adding votersDV – deleting votersPV – partition of voters
Results: Control
Llull
/ Copeland
0
Plurality
Control
CC
DC
CC
DC
AC
R
V
R
R
DC
R
V
R
R
(R)PC-TP
R
V
R
R
(R)PC-TE
R
V
R
R
PV-TP
R
R
V
V
PV-TE
R
R
R
R
AV
R
R
V
V
DV
R
R
V
V
CC
– constructive control
TP
– ties promote
DC
– destructive control
TE
– ties eliminate
R
– NP-complete
V
– P membershipSlide36
Outline
IntroductionComputational study of electionsBribery and control
Llull/Copeland ElectionsModel of electionsRepresentation of votes Llull/Copeland ruleResultsBribery and microbribery
Control of electionsManipulation
My coalition of agents must be able to somehow manipulate the system!!Slide37
Manipulation
Manipulation problemGiven: Set of candidates Set of honest voters (with their preferences)
Set of manipulatorsPreferred candidate pQuestion: Can the manipulators set their votes so as to guarantee p’s victory?Can’t avoid manipulation!
Gibbard-SatterthwaiteDugan-Schwarz
There Llull! By Gibbard-Satterthwaite result I know I can sometimes manipulate your system. Finally math gets useful!Slide38
Results: Manipulation
Unweighted manipulationNP-complete for all rational tie-handling values except
0, 0.5, 1Even for two manipulators!Proof ideasAlmost a microbribery instance
The two manipulators can only switch head-to-head results from tie to a victoryReduce from X3C,1-in-3-SAT
Can’t use microbribery to solve the problem because the two voters might not be able to implement the solution we would get
Haha Mr. Agent! Gibbard-Satterthwaite say that you may be able to influence my system, but you won’t know when or how!Slide39
Results: Manipulation
Unweighted manipulationNP-complete for all rational tie-handling values except
0, 0.5, 1Even for two manipulators!Aren’t we missing exactly the interesting values?What is magic about the values we miss?
1 no point in causing ties0.5 switching the result of a head-to-head contest from tie to win affect both candidates in a symmetric way
0 makes “transferring points” difficult
So… maybe there is hope for me…Slide40
Results: Manipulation
Weighted manipulationCompletely resolved for three candidatesIssues with for and more candidates
Unique vs. non-unique winnerTie-value 1 is hard to account for.Strange things happen for four candidates.Weighted, three candidates
= 0
0<<1
= 1
winner
NPC
NPC
P
unique
winner
NPC
P
P
Work in progressSlide41
Summary
Copeland/Llull electionsBroad resistance to briberyConstructive/DestructiveRational/Irrational
Broad resistance to controlConstructive candidate controlConstructive/Destructive voter controlRational/IrrationalVulnerabilityMicrobribery (irrational case)
Arrgh! Llull, my agents are practically helpless against your system!Slide42
Thank you!
I will happily answer your questions!