FIND ME ON

GitHub

LinkedIn

Information Function

🌱

Definition
StochasticControl

Definition

Static

For a team of NN agents, for each agent, Ai,iN\mathbf{A} i,i\in\mathcal{N} the information function, ηi\eta^{i}, maps from some standard Borel space (Ξ,B(Ξ))(\mathit\Xi,\mathcal{B}(\mathit{\Xi})) to the observable space (Yi,Yi)(\mathbb{Y}^{i},\mathcal{Y}^{i}) (defined in Witsenhausen’s Intrinsic Model): ηi:(Ξ,B(Ξ))(Yi,Yi)\eta^{i}:(\mathit\Xi,\mathcal{B}(\mathit\Xi))\to(\mathbb{Y}^{i},\mathcal{Y}^{i})by convention though, we can treat this standard Borel space as being the same as the probability space (Ω,F)(\Omega,\mathcal{F}) hence ηi:(Ω,F)(Yi,Yi)\eta^{i}:(\Omega,\mathcal{F})\to(\mathbb{Y}^{i},\mathcal{Y}^{i}) ## Dynamic In the dynamic case we take past actions into account as well hence ηi:(Ω,F)×(Ut1,Ut1)××(UtN,UtN)(Yi,Yi),tT,iN\eta^{i}:(\Omega,\mathcal{F})\times (\mathbb{U}^{1}_{t},\mathcal{U}^{1}_{t})\times\dots \times (\mathbb{U}^{N}_{t},\mathcal{U}^{N}_{t})\to(\mathbb{Y}^{i},\mathcal{Y}^{i}),\quad t\in\mathcal{T}, i\in\mathcal{N}

Linked from