1 Evolution & Economics No. 5 Forest fire p = 0.53 p = 0.58.
-
Upload
theresa-harrington -
Category
Documents
-
view
218 -
download
1
Transcript of 1 Evolution & Economics No. 5 Forest fire p = 0.53 p = 0.58.
1
Evolution & EconomicsEvolution & EconomicsNo. 5No. 5
Forest fire
p = 0.53 p = 0.58
2
Best Response DynamicsBest Response Dynamics
• A finite population playing a (symmetric) game G.
• Each individual is randomly matched and plays a pure strategy of G.
• At each point in (discrete) time each individual (or alternatively each with probability) chooses the
best response to the mixed strategy played by the population.
3
A B
A 3 , 3 0 , x
B x , 0 x , x
Payoff Dominant EquilibriumA B
A 3 , 3 0 , x
B x , 0 x , x
3 > x > 0
Nash EquilibriaNash Equilibria
When then: 3+ 0x >
2
Risk Dominant Equilibrium
4
Payoff Dominant Equilibrium
Risk Dominant Equilibrium
A B
A 3 , 3 0 , x
B x , 0 x , x
A Population of n individuals playing either A or B.
A B
All playing A All playing B
0 1k/n
k playing B
5
Payoff Dominant Equilibrium
Risk Dominant Equilibrium
A B
A 3 , 3 0 , x
B x , 0 x , x
A B
A 3 , 3 0 , 2
B 2 , 0 2 , 2
2
222
1-α α Best Response to is:
when when
1- α,α
1 1A > α, B < α.3 3
A B13
6
Payoff Dominant Equilibrium
Risk Dominant Equilibrium
A B
A 3 , 3 0 , x
B x , 0 x , x
A B
A 3 , 3 0 , 2
B 2 , 0 2 , 2
2
222
A B13
In the Best Response Dynamics:In the Best Response Dynamics:The The Basin of AttractionBasin of Attraction of the Risk Dominant Equilibrium of the Risk Dominant Equilibrium
is is largerlarger than that of the Pareto Dominant Equilibrium than that of the Pareto Dominant Equilibrium
7
Let a state of the population be a complete description
of the population:
ns Rwho is playing what.
011s =.........
A
B
There are states of the population.n2
8
Let a state of the population be a complete description
of the population:
ns Rwho is playing what.
011s =.........
A
B
There are states of the population.n2
A deterministic dynamics can be described as a transition functionA deterministic dynamics can be described as a transition function
.i js s
9
The replicator Dynamics The replicator Dynamics (in which all learn – revise their strategy) is:(in which all learn – revise their strategy) is:
is 0 or 1
If each individual has probabilityIf each individual has probability λ of learning: of learning:
A B13
10If each individual has probabilityIf each individual has probability λ of learning: of learning:
A B13
A state of the population becomes a distribution of statesA state of the population becomes a distribution of states
11
The prob. of learning may be dropped.The prob. of learning may be dropped.
A B13
If each individual (when learning) has a small probabilityIf each individual (when learning) has a small probability εof learning the ‘wrong’ strategy:of learning the ‘wrong’ strategy:
B (B)
B (A)
Each state has a positive probability of becoming any other state.Each state has a positive probability of becoming any other state.
12
Let be a distribution over the states of the population.n p 2
Let be the probability that state becomes
jij
ji i j(ε) > 0
t = 1.
t s s
,Let be the transition matrix.
describes a
ji
tt+1
ε
T = t
p = Tp Markov process.
lim
is independent of
the columns of are identical.
t 00
t
tt 0
p p
T
p = T
T
p
13
lim..
n n n
1 1 1
2 2 2tt
2 2 2
p p .. pp p p
T = T = .. .. ..
p p .. p
lim t 0 0
t p = T p = pT lim t
tε ε0 0ε ε p = T p = pT
lim
What is the limit distribution when
ε 0ε
ε 0 ????
p = ???
14
A B13
A B
A 3 , 3 0 , 2
B 2 , 0 2 , 2
ε-small, mistakes are rare
The process converges to B (A) and will stay there a long time.
ε 0
In the long-run mistakes occur (with prob. 1) and the process will be swept away from B.
15
A B13
If more individuals made mistakes, the process moves to the basin of attraction of A, from there to A, and will remain there a long time.
ε 0
If few individuals make mistakes, the process will remain in the basin of attraction of B, and will return to B.
16
A B13
What proportion of the time is spent in A ??
ε 0
The process remains a long time in A and in B.
To leave the basin of attraction of B, at least 2/3 of the population has to err. This happens with prob.:
n
k
2k= n3
2n3n
εk
ε
To leave the basin of attraction of A, at least 1/3 of the population has to err. This happens with prob. :
n
k
1k= n3
1n3n
εk
ε
17
A B13
ε 0
1n3ε
22 1n n3 3=ε ε
It is a rare move from B to the basin of attraction of A.The reverse is much more frequent !!!
1n
13
3
3
n2n
leaving A
leaving =
B= εε
ε0
18
A B13
ε 0
1n3ε
22 1n n3 3=ε ε
In the very long run the process will spend infinitely more time in B than in A.
1n
13
3
3
n2n
leaving A
leaving =
B= εε
ε0
In the very long run the process will spend 0 time in A.
19
The Best Response Dynamics – with vanishing noise, chooses the Risk Dominant Equilibrium. The equilibrium with a larger Basin of attraction.
A B
A 3 , 3 0 , 2
B 2 , 0 2 , 2
A B
A 3 , 3 0 , 2
B 2 , 0 2 , 2
20
Kandori, Michihiro & Mailath, George J & Rob, Rafael, 1993. "Learning, Mutation, and Long Run Equilibria in Games," Econometrica, Econometric Society, vol. 61(1), pages 29-56.
H Peyton, 1993. "The Evolution of Conventions," Econometrica, EconometricSociety, vol. 61(1), pages 57-84
M.I. Freidlin and A.D. Wentzell, Random Perturbations of Dynamical Systems , Springer New York, 1984
Based on: