Next-generation matrix and the basic reproduction number
DefinitionBasic reproduction number of the next-generation matrix#✓fully proved
The basic reproduction number R₀ = ρ(F V⁻¹): the spectral radius (largest modulus of
a complex eigenvalue) of the next-generation matrix F V⁻¹ (van den Driessche–Watmough 2002,
Eq. (4)). It is meaningful when V is invertible (for a nonsingular M-matrix it is).
noncomputable def R0 (F V : Matrix ι ι ℝ) : ℝ :=
(spectralRadius ℂ ((F * V⁻¹).map (algebraMap ℝ ℂ))).toReal
TheoremR₀ as a threshold for the linearized infected subsystem#✓fully proved
R₀ as a threshold for the linearized infected subsystem (van den Driessche–Watmough
2002, proof of Theorem 2). If F is entrywise nonnegative and V is a nonsingular M-matrix,
then every eigenvalue of F - V has negative real part iff R₀ = ρ(F V⁻¹) < 1.
theorem isHurwitz_sub_iff_R0_lt_one {F V : Matrix ι ι ℝ} (hF : ∀ i j, 0 ≤ F i j)
(hV : V.IsNonsingularMMatrix) : (F - V).IsHurwitz ↔ R0 F V < 1
noncomputable def R0 (F V : Matrix ι ι ℝ) : ℝ :=
(spectralRadius ℂ ((F * V⁻¹).map (algebraMap ℝ ℂ))).toReal
MathlibExt/Biology/Epidemiology/NextGenerationMatrix.lean:257
Uses 7 · Used by 1
TheoremR₀ > 1 gives an unstable mode#✓fully proved
R₀ > 1 gives an unstable mode (van den Driessche–Watmough 2002, proof of Theorem 2).
If F is entrywise nonnegative and V is a nonsingular M-matrix, then F - V has an
eigenvalue with positive real part iff R₀ = ρ(F V⁻¹) > 1.
theorem exists_re_pos_iff_one_lt_R0 {F V : Matrix ι ι ℝ} (hF : ∀ i j, 0 ≤ F i j)
(hV : V.IsNonsingularMMatrix) :
(∃ μ ∈ spectrum ℂ ((F - V).map (algebraMap ℝ ℂ)), 0 < μ.re) ↔ 1 < R0 F V
noncomputable def R0 (F V : Matrix ι ι ℝ) : ℝ :=
(spectralRadius ℂ ((F * V⁻¹).map (algebraMap ℝ ℂ))).toReal
MathlibExt/Biology/Epidemiology/NextGenerationMatrix.lean:257
Uses 7 · Used by 1
TheoremEigenvalues of a block lower-triangular matrix#✓fully proved
The complex eigenvalues of a block lower-triangular real matrix [[A, 0], [C, D]] are
those of A together with those of D.
theorem mem_spectrum_fromBlocks_zero₁₂ {A : Matrix ι ι ℝ} {C : Matrix κ ι ℝ}
{D : Matrix κ κ ℝ} {μ : ℂ} :
μ ∈ spectrum ℂ ((fromBlocks A 0 C D).map (algebraMap ℝ ℂ)) ↔
μ ∈ spectrum ℂ (A.map (algebraMap ℝ ℂ)) ∨ μ ∈ spectrum ℂ (D.map (algebraMap ℝ ℂ))
TheoremHurwitz criterion for block lower-triangular matrices#✓fully proved
A block lower-triangular real matrix [[A, 0], [C, D]] is Hurwitz iff both diagonal
blocks are.
theorem isHurwitz_fromBlocks_zero₁₂_iff {A : Matrix ι ι ℝ} {C : Matrix κ ι ℝ}
{D : Matrix κ κ ℝ} : (fromBlocks A 0 C D).IsHurwitz ↔ A.IsHurwitz ∧ D.IsHurwitz
Uses 2 · Used by 1
TheoremThreshold theorem, stability half, conditional decay form#✓fully proved
Threshold theorem, stability half, conditional decay form (van den Driessche–Watmough
2002, Theorem 2). Let x₀ be an equilibrium of x' = f(x) on ℝ^{ι ⊕ κ} (infected
compartments ι, the others κ), with f differentiable at x₀ and Jacobian
[[F - V, 0], [C, D]] as in their Lemma 1, where F ≥ 0, V is a nonsingular M-matrix and
D is Hurwitz. If R₀ = ρ(F V⁻¹) < 1, then every solution defined on all of [0, ∞)
starting near x₀ decays exponentially. The statement concerns only solutions defined on all of [0, ∞), so it is not by itself local exponential stability.
theorem globalSolution_decay_of_R0_lt_one {f : (ι ⊕ κ → ℝ) → ι ⊕ κ → ℝ}
{x₀ : ι ⊕ κ → ℝ} {F V : Matrix ι ι ℝ} {C : Matrix κ ι ℝ} {D : Matrix κ κ ℝ}
(hf : f x₀ = 0)
(hfJ : HasFDerivAt f
(LinearMap.toContinuousLinearMap (Matrix.toLin' (fromBlocks (F - V) 0 C D))) x₀)
(hF : ∀ i j, 0 ≤ F i j) (hV : V.IsNonsingularMMatrix) (hD : D.IsHurwitz)
(hR0 : R0 F V < 1) :
∃ δ K α : ℝ, 0 < δ ∧ 0 < K ∧ 0 < α ∧ ∀ x : ℝ → ι ⊕ κ → ℝ,
(∀ t, 0 ≤ t → HasDerivAt x (f (x t)) t) → ‖x 0 - x₀‖ < δ →
∀ t, 0 ≤ t → ‖x t - x₀‖ ≤ K * Real.exp (-(α * t)) * ‖x 0 - x₀‖
noncomputable def R0 (F V : Matrix ι ι ℝ) : ℝ :=
(spectralRadius ℂ ((F * V⁻¹).map (algebraMap ℝ ℂ))).toReal
MathlibExt/Biology/Epidemiology/NextGenerationMatrix.lean:257
Uses 6
TheoremThreshold theorem, instability half, spectral form#✓fully proved
Threshold theorem, instability half, spectral form (van den Driessche–Watmough 2002,
Theorem 2). With F ≥ 0 and V a nonsingular M-matrix, if R₀ = ρ(F V⁻¹) > 1 then the
Jacobian [[F - V, 0], [C, D]] at the disease-free equilibrium has an eigenvalue with positive
real part (whatever C and D).
theorem exists_re_pos_jacobian_of_one_lt_R0 {F V : Matrix ι ι ℝ} (C : Matrix κ ι ℝ)
(D : Matrix κ κ ℝ) (hF : ∀ i j, 0 ≤ F i j) (hV : V.IsNonsingularMMatrix)
(hR0 : 1 < R0 F V) :
∃ μ ∈ spectrum ℂ ((fromBlocks (F - V) 0 C D).map (algebraMap ℝ ℂ)), 0 < μ.re
noncomputable def R0 (F V : Matrix ι ι ℝ) : ℝ :=
(spectralRadius ℂ ((F * V⁻¹).map (algebraMap ℝ ℂ))).toReal
MathlibExt/Biology/Epidemiology/NextGenerationMatrix.lean:257
Uses 4
Loading dependency explorer…
The dependency explorer needs JavaScript and its graph data.