Title: An introduction to game dynamics
1An introduction to game dynamics
2(No Transcript)
3Nash equilibria
4Nash equilibria
5Nash equilibria
6 - The greatest conceptual revolution in
biology...the replacement of typological
thinking by population thinking. -
- Ernst Mayr
7John Maynard Smith
8Symmetric Games
9Chicken Game
10Population setting
11Population setting
12Replicator equation
13Replicator dynamics and Nash equilibria
14Folk theorem of evolutionary game theory
- Nash equilibria are rest points
- strict Nash equilibria are attractors
- stable rest points are Nash equilibria
- limits of interior orbits are Nash equilibria
15Existence of Nash equilibria
16Existence of Nash equilibria
17Existence of Nash equilibria
18Existence of Nash equilibria
19Existence of Nash equilibria
20Replicator equation for n2
21Replicator equation for n2
- Dominance
- Bistability
- stable coexistence
22Example dominance
23Example bistability
24Example coexistence
25Replicator equations for n3
- Possibility for
- rock-scissors-paper
- heteroclinic cycles
26 Rock -Paper- Scissors
for the adult player
27(No Transcript)
28(No Transcript)
29(No Transcript)
30Rock-Paper-Scissors
31Rock-Paper-Scissors
32(No Transcript)
33Generalized Rock-Paper-Scissors
34Generalized Rock-Paper-Scissors
35Generalized Rock-Paper-Scissors
36Generalized Rock-Paper-Scissors
37Generalized Rock-Paper-Scissors
38Proof (Zeeman, Hofbauer,)
39 40 41In the Stone Age
42Bacterial Game Dynamics
- Escherichia coli
- Type A wild type
- Type B mutant producing colicin (toxic) and an
immunity protein - Type C produces only the immunity protein
43Bacterial Game Dynamics
- Escherichia coli
- Rock-Paper-Scissors cycle
- Not permanent!
- Serial transfer (from flask to flask)
- only one type can survive!
- (Kerr et al, Nature 2002)
44Mating behavior
- Uta stansburiana (lizards)
- (Sinervo and Lively, Nature, 1998)
45(No Transcript)
46Mating behavior
-
- males 3 morphs (inheritable)
47(No Transcript)
48Rock-Paper-Scissors in Nature
- males 3 morphs (inheritable)
- A monogamous, guards female
- B polygamous, guards harem (less efficiently)
- C loose males, sneaky matings
49Other examples of RPS-cycles
In models of ecological competition
(May,) Non-permanent community can be
stabilised by fourth population (competitor or
predator) (Hofbauer, Schreiber,)
50Other examples of RPS-cycles
- In the evolution of virulence
- Three strains of a pathogen
- Different virulence (extra mortality)
- Super-infection of host organism
(most virulent takes over) - Different infectivity
- (Nowak, May,)
51Other examples of RPS-cycles
- In genetics
- Segregation distorters (outlaw genes)
- Harm their opposite chromosome before gametes are
formed - (Crow, Grafen, Stadler)
52Iterated Prisoners Dilemma
- infinitely repeated, with errors
- All D
- Tit For Tat
- Firm But Fair
53Iterated Prisoners Dilemma
Heteroclinic network A Tit For Tat B Firm But
Fair C Bully D AllD
54Other game dynamics
- Transmission of strategies through
inheritance, imitation, learning
55Imitation dynamics
56Imitation dynamics
57Imitate the better
58Best Reply Dynamics
59Best Reply
60Best reply dynamics
61Best Reply
62Best Reply
63Are there game dynamics which always converge
to Nash equilibria?
64Myopic adjustment (Hofbauer)
65Myopic adjustment (Hofbauer)
66Phase portraits of Replicator equations
67Public Goods
- all do better if there are more cooperators in
the group - within each group, defectors do better than
cooperators - temptation to free-ride and exploit co-players
- no public good!
68Public Goods Experiments
- 6 players
- 1 dollar each
- contribute to common pool?
- Experimenter triples amount in pool
- divides it equally among all players
69Public Goods Experiments
- 6 players
- 1 dollar each
- contribute to common pool?
- Experimenter triples amount in pool
- divides it equally among all players
- return for each dollar only 50 cents!
- not to contribute is dominant
70Optional Public Goods
- (Hauert et al, Science 2002)
- large population
- N players offered to participate (sample)
- S accept (group)
- N-S decline (loners, risk-averse)
- loners have fallback-solution
71Optional Public Goods
- three strategies
- loner
- cooperator
- defector
- (if only one participates loner)
72Game Dynamics
73Payoff for optional game
74Optional Public Goods
75Replicator Dynamics
76Rock-Paper-Scissors cycle
- If most cooperate, best to defect
- if most defect, best to abstain
- if most abstain, best to cooperate
- (for small groups, Simpsons paradox)
77Simpsons paradox
78Change in variables
79Hamiltonian system
80Replicator dynamics
81Time averages
82Red Queen Dynamics
83Optional Public Goods
- Well-mixed population (random encounters)
84Spatial Games
- Interaction with nearest neighbors only
- best takes over
85Optional Public Goods
86looking back to bacterial games
- Rock-Paper-Scissors cycle
- Not permanent!
- Serial transfer (from flask to flask)
- only one type can survive!
87looking back to bacterial games
- Rock-Paper-Scissors cycle
- Not permanent!
- Serial transfer (from flask to flask)
- only one type can survive!
- Serial transfer from plate to plate
- all types survive!
- (Kerr et al, Nature 2002)
88Asymmetric Games
89Asymmetric Games
- Owner vs. Intruder
- Female vs. Male
- Young vs. Old
- Buyer vs. Seller
- Host vs. Parasite
90Male vs. Female Battle of the Sexes
- Males invest little
- Females invest much
- Females must be choosy
- long engagement periods?
- Females coy or fast
- Males faithful or philandering
91Battle of the Sexes
- Females coy
- Males faithful
- Females fast
- Males philanderers
- Females coy
- ...
92Battle of the Sexes
93Battle of the Sexes
94Battle of the Sexes
95Example Owner-Intruder
96Example Owner-Intruder
97Example Owner-Intruder
98Example Owner-Intruder
- (1,0) Bourgeois strategy
- (0,1) paradox strategy (Prudhon)
- asymmetric Nash equilibria
99Conditional Strategies
- If owner, escalate if intruder, display
- If male, philanderer if female, coy
100Conditional Strategies
101Conditional Strategies
102Conditional Strategies
103Conditional Strategies
104Punishment
- after public goods game, players can fine
co-players - punishment is costly
- punishment is efficient
- works even if players never meet again!
- (Fehr and Gächter, 2002)
105(No Transcript)
106(No Transcript)
107Punishment
- introduces a
- second order social dilemma!
108Evolutionary Game Dynamics
109Strategies for Public Goods with Punishment
110Payoff for Public Goods with Punishment
111 112 113 Public Goods with Punishment
114 Public Goods with Punishment
115Reputation and Temptation
116additional payoff terms
117Reputation effect
118Bifurcation
119Morals?
- More freedom yields more cooperation
- Individuals that are less social
- make better societies
120 - Christoph Hauert
- Josef Hofbauer
- Martin Nowak
- Silvia Demonte
- Hannelore Brandt