TheoremBase

The Transition-Rate Family and Aggregate Drift of Affine-Controlled Data

Statement

Let (β0,β1)(\beta_0,\beta_1) be an affine-controlled transition-rate family on ll states with control set A\mathcal{A} and Lipschitz constant Λ\Lambda, and let Δl\Delta^l be the probability simplex. Define

β(σ,γ,Σ,α)=β0(σ,γ,Σ)+β1(σ,γ,Σ)⋅α(Σ∈Δl, α∈A),\beta(\sigma,\gamma,\Sigma,\alpha)=\beta_0(\sigma,\gamma,\Sigma)+\beta_1(\sigma,\gamma,\Sigma)\cdot\alpha\qquad(\Sigma\in\Delta^l,\ \alpha\in\mathcal{A}),

for each ordered pair (σ,γ)(\sigma,\gamma) with σ≠γ\sigma\neq\gamma in {1,…,l}\{1,\dots,l\}, where x⋅yx\cdot y denotes the dot product and ∣⋅∣|\cdot| the Euclidean norm. Let

R=sup⁡α∈A∣α∣,K1=sup⁡∣β1(σ,γ,Σ)∣,B=sup⁡β(σ,γ,Σ,α),R=\sup_{\alpha\in\mathcal{A}}|\alpha|,\qquad K_1=\sup|\beta_1(\sigma,\gamma,\Sigma)|,\qquad B=\sup\beta(\sigma,\gamma,\Sigma,\alpha),

the latter suprema being over all ordered pairs (σ,γ)(\sigma,\gamma) with σ≠γ\sigma\neq\gamma, all Σ∈Δl\Sigma\in\Delta^l and all α∈A\alpha\in\mathcal{A}, and set Λβ=Λ(1+R)\Lambda_\beta=\Lambda(1+R) and Λb=2l (l−1)(B+Λβ)\Lambda_b=2\sqrt{l}\,(l-1)(B+\Lambda_\beta).

1. (Finiteness.) RR, K1K_1 and BB are finite nonnegative real numbers, and the nearest-point projection πΔl\pi_{\Delta^l} onto Δl\Delta^l is defined.

2. (Transition-rate family.) β\beta is a transition-rate family on ll states with control set A\mathcal{A} and rate bound BB. Moreover, for each fixed α∈A\alpha\in\mathcal{A} the map Σ↦β(σ,γ,Σ,α)\Sigma\mapsto\beta(\sigma,\gamma,\Sigma,\alpha) is Lipschitz with constant Λβ\Lambda_\beta on Δl\Delta^l.

3. (Affine drift.) Let bb be the aggregate state drift of β\beta, and define, for Σ∈Δl\Sigma\in\Delta^l and γ∈{1,…,l}\gamma\in\{1,\dots,l\},

b0γ(Σ)=∑σ≠γ(Σσβ0(σ,γ,Σ)−Σγβ0(γ,σ,Σ)),b1γ(Σ)=∑σ≠γ(Σσβ1(σ,γ,Σ)−Σγβ1(γ,σ,Σ)),b^\gamma_0(\Sigma)=\sum_{\sigma\neq\gamma}\big(\Sigma^\sigma\beta_0(\sigma,\gamma,\Sigma)-\Sigma^\gamma\beta_0(\gamma,\sigma,\Sigma)\big),\qquad b^\gamma_1(\Sigma)=\sum_{\sigma\neq\gamma}\big(\Sigma^\sigma\beta_1(\sigma,\gamma,\Sigma)-\Sigma^\gamma\beta_1(\gamma,\sigma,\Sigma)\big),

with values in R\mathbb{R} and Rm\mathbb{R}^m respectively. Then

bγ(Σ,α)=b0γ(Σ)+b1γ(Σ)⋅αfor all Σ∈Δl, α∈A.b^\gamma(\Sigma,\alpha)=b^\gamma_0(\Sigma)+b^\gamma_1(\Sigma)\cdot\alpha\qquad\text{for all }\Sigma\in\Delta^l,\ \alpha\in\mathcal{A}.

4. (Bounds.) For all Σ,Σ′∈Δl\Sigma,\Sigma'\in\Delta^l and all α,α′∈A\alpha,\alpha'\in\mathcal{A},

∣b(Σ,α)∣≤2l (l−1)B,∣b(Σ,α)−b(Σ′,α)∣≤Λb ∣Σ−Σ′∣,∣b(Σ,α)−b(Σ,α′)∣≤2l (l−1)K1∣α−α′∣.|b(\Sigma,\alpha)|\le2\sqrt{l}\,(l-1)B,\qquad |b(\Sigma,\alpha)-b(\Sigma',\alpha)|\le\Lambda_b\,|\Sigma-\Sigma'|,\qquad |b(\Sigma,\alpha)-b(\Sigma,\alpha')|\le2\sqrt{l}\,(l-1)K_1|\alpha-\alpha'| .

5. (Conservation and inflow bound.) For all Σ∈Δl\Sigma\in\Delta^l, α∈A\alpha\in\mathcal{A} and γ∈{1,…,l}\gamma\in\{1,\dots,l\},

∑γ=1lbγ(Σ,α)=0,bγ(Σ,α)≥−(l−1)B Σγ.\sum_{\gamma=1}^lb^\gamma(\Sigma,\alpha)=0,\qquad b^\gamma(\Sigma,\alpha)\ge-(l-1)B\,\Sigma^\gamma .

6. (Projected drift.) Define b^:Rl×A→Rl\hat{b}:\mathbb{R}^l\times\mathcal{A}\to\mathbb{R}^l by b^(x,α)=b(πΔl(x),α)\hat{b}(x,\alpha)=b(\pi_{\Delta^l}(x),\alpha). Then b^(x,α)=b(x,α)\hat{b}(x,\alpha)=b(x,\alpha) for x∈Δlx\in\Delta^l, ∣b^(x,α)∣≤2l (l−1)B|\hat{b}(x,\alpha)|\le2\sqrt{l}\,(l-1)B for all x∈Rlx\in\mathbb{R}^l and α∈A\alpha\in\mathcal{A}, and ∣b^(x,α)−b^(x′,α)∣≤Λb∣x−x′∣|\hat{b}(x,\alpha)-\hat{b}(x',\alpha)|\le\Lambda_b|x-x'| for all x,x′∈Rlx,x'\in\mathbb{R}^l and α∈A\alpha\in\mathcal{A}.

Proofs

Log in to submit a proof.

Loading...

Citations

Loading…

Dependencies

Loading…

Related

0 relations

Curated associations between results. These are editable and subjective — they do not replace the dependency graph, which is derived from the references in the text.

No relations recorded yet.

Comments

Log in to comment.

Loading…