Replicator dynamics
Replicator
DefinitionMean fitness of a population state#✓fully proved
The mean fitness f̄(x) = ∑ i, x i * f x i of the population state x for the fitness
function f.
def meanFitness (f : (ι → ℝ) → ι → ℝ) (x : ι → ℝ) : ℝ :=
∑ i, x i * f x i
Used by 7
StructureSolution of the replicator equation#✓fully proved
x solves the replicator equation x_i' = x_i (f_i(x) - f̄(x)) for all times t ≥ 0.
structure IsSolution (f : (ι → ℝ) → ι → ℝ) (x : ℝ → ι → ℝ) : Prop where
/-- Each frequency grows at the rate of its excess fitness. -/
hasDerivAt : ∀ i t, 0 ≤ t →
HasDerivAt (fun s => x s i) (x t i * (f (x t) i - meanFitness f (x t))) t
def meanFitness (f : (ι → ℝ) → ι → ℝ) (x : ι → ℝ) : ℝ :=
∑ i, x i * f x i
MathlibExt/Biology/Dynamics/Replicator.lean:61
Uses 1 · Used by 5
DefinitionRest point of the replicator equation#✓fully proved
x is a rest point of the replicator equation: x_i (f_i(x) - f̄(x)) = 0 for all i.
def IsRestPoint (f : (ι → ℝ) → ι → ℝ) (x : ι → ℝ) : Prop :=
∀ i, x i * (f x i - meanFitness f x) = 0
def meanFitness (f : (ι → ℝ) → ι → ℝ) (x : ι → ℝ) : ℝ :=
∑ i, x i * f x i
MathlibExt/Biology/Dynamics/Replicator.lean:61
Uses 1 · Used by 1
TheoremRest points in the interior are equalizers#✓fully proved
If every strategy is present (x i > 0),
then x is a rest point iff every strategy has fitness equal to the mean fitness.
theorem isRestPoint_iff {f : (ι → ℝ) → ι → ℝ} {x : ι → ℝ} (hx : ∀ i, 0 < x i) :
IsRestPoint f x ↔ ∀ i, f x i = meanFitness f x
def IsRestPoint (f : (ι → ℝ) → ι → ℝ) (x : ι → ℝ) : Prop :=
∀ i, x i * (f x i - meanFitness f x) = 0
MathlibExt/Biology/Dynamics/Replicator.lean:71
def meanFitness (f : (ι → ℝ) → ι → ℝ) (x : ι → ℝ) : ℝ :=
∑ i, x i * f x i
MathlibExt/Biology/Dynamics/Replicator.lean:61
Uses 2
Is Solution
TheoremEvolution of the total frequency under replicator dynamics#✓fully proved
The total frequency ∑ x_i evolves by (∑ x_i)' = f̄(x) (1 - ∑ x_i).
theorem hasDerivAt_sum (h : IsSolution f x) {t : ℝ} (ht : 0 ≤ t) :
HasDerivAt (fun s => ∑ i, x s i) (meanFitness f (x t) * (1 - ∑ i, x t i)) t
structure IsSolution (f : (ι → ℝ) → ι → ℝ) (x : ℝ → ι → ℝ) : Prop where
/-- Each frequency grows at the rate of its excess fitness. -/
hasDerivAt : ∀ i t, 0 ≤ t →
HasDerivAt (fun s => x s i) (x t i * (f (x t) i - meanFitness f (x t))) t
MathlibExt/Biology/Dynamics/Replicator.lean:65
def meanFitness (f : (ι → ℝ) → ι → ℝ) (x : ι → ℝ) : ℝ :=
∑ i, x i * f x i
MathlibExt/Biology/Dynamics/Replicator.lean:61
Uses 2 · Used by 1
TheoremInvariance of the simplex hyperplane#✓fully proved
For continuous fitness f, a solution with
∑ x_i(0) = 1 satisfies ∑ x_i(t) = 1 for all t ≥ 0.
theorem sum_eq_one (h : IsSolution f x) (hf : Continuous f) (h0 : ∑ i, x 0 i = 1) {t : ℝ}
(ht : 0 ≤ t) : ∑ i, x t i = 1
structure IsSolution (f : (ι → ℝ) → ι → ℝ) (x : ℝ → ι → ℝ) : Prop where
/-- Each frequency grows at the rate of its excess fitness. -/
hasDerivAt : ∀ i t, 0 ≤ t →
HasDerivAt (fun s => x s i) (x t i * (f (x t) i - meanFitness f (x t))) t
MathlibExt/Biology/Dynamics/Replicator.lean:65
Uses 3 · Used by 1
TheoremInvariance of the faces of the simplex#✓fully proved
For continuous fitness f, a strategy absent
at time 0 (x_i(0) = 0) stays absent for all t ≥ 0.
theorem eq_zero_of_eq_zero (h : IsSolution f x) (hf : Continuous f) {i : ι} (h0 : x 0 i = 0)
{t : ℝ} (ht : 0 ≤ t) : x t i = 0
structure IsSolution (f : (ι → ℝ) → ι → ℝ) (x : ℝ → ι → ℝ) : Prop where
/-- Each frequency grows at the rate of its excess fitness. -/
hasDerivAt : ∀ i t, 0 ≤ t →
HasDerivAt (fun s => x s i) (x t i * (f (x t) i - meanFitness f (x t))) t
MathlibExt/Biology/Dynamics/Replicator.lean:65
Uses 2 · Used by 1
TheoremPresent strategies stay present#✓fully proved
For continuous fitness f, a strategy with positive
frequency at time 0 has positive frequency for all t ≥ 0.
theorem pos_of_pos (h : IsSolution f x) (hf : Continuous f) {i : ι} (h0 : 0 < x 0 i)
{t : ℝ} (ht : 0 ≤ t) : 0 < x t i
structure IsSolution (f : (ι → ℝ) → ι → ℝ) (x : ℝ → ι → ℝ) : Prop where
/-- Each frequency grows at the rate of its excess fitness. -/
hasDerivAt : ∀ i t, 0 ≤ t →
HasDerivAt (fun s => x s i) (x t i * (f (x t) i - meanFitness f (x t))) t
MathlibExt/Biology/Dynamics/Replicator.lean:65
Uses 2 · Used by 1
TheoremInvariance of the simplex (Hofbauer–Sigmund 1998, §7.1)#✓fully proved
For continuous fitness f, a
solution starting in the standard simplex (x i ≥ 0 for all i and ∑ i, x i = 1) stays in
it for all t ≥ 0.
theorem mem_simplex (h : IsSolution f x) (hf : Continuous f) (h0 : ∀ i, 0 ≤ x 0 i)
(h1 : ∑ i, x 0 i = 1) {t : ℝ} (ht : 0 ≤ t) : (∀ i, 0 ≤ x t i) ∧ ∑ i, x t i = 1
structure IsSolution (f : (ι → ℝ) → ι → ℝ) (x : ℝ → ι → ℝ) : Prop where
/-- Each frequency grows at the rate of its excess fitness. -/
hasDerivAt : ∀ i t, 0 ≤ t →
HasDerivAt (fun s => x s i) (x t i * (f (x t) i - meanFitness f (x t))) t
MathlibExt/Biology/Dynamics/Replicator.lean:65
Uses 4
Loading dependency explorer…
The dependency explorer needs JavaScript and its graph data.