Invariance of stochastic control systems with deterministic arguments (Q1880747)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Invariance of stochastic control systems with deterministic arguments
scientific article

    Statements

    Invariance of stochastic control systems with deterministic arguments (English)
    0 references
    0 references
    0 references
    1 October 2004
    0 references
    The authors prove that a closed set \(K\) of a finite-dimensional space is invariant under the stochastic control system \[ dX =b(X,v(t))\,dt+\sigma(x,v(t))\,dW(t),\quad v(t)\in U, \] if and only if it is invariant under the deterministic control system with two controls \[ x'=b(x,v(t))-\frac{1}{2}\sum_{j=1}^mD\,\sigma_j(x,v(t))\sigma_j(x,v(t))+ \sigma(x,v(t))u(t),\,\, u(t)\in H_1,\,\,v(t)\in U. \] This extends the results of \textit{H. Doss} [``Liens entre équations différentielles stochastiques et ordinaires'', Ann. Inst. Henri Poincaré, Nouv. Sér., Sect. B. 13, 99--125 (1977; Zbl 0359.60087)] into two directions: to control systems and less regular \(\sigma\). Only \(C^{1,1}\) regularity of the diffusion \(\sigma\) is asked for instead of the usual assumption \(\sigma\in C^2\). A very direct `deterministic proof' is proposed, while the Doss arguments are based on the support theorem of stochastic analysis. In this way this result is new even for stochastic differential equations. The arguments of the proof are based on estimates between solutions of the stochastic control system with time-independent controls and families of solutions \(\{x_\omega(\cdot)\}_{\omega\in\Omega}\) to the deterministic control system \[ x'=\sigma(x,v_\omega)u_\omega(t),\quad u_\omega(t)\in H_1 \] with appropriately chosen controls \(u_\omega(t)\) and \(v_\omega(t)\in U\).
    0 references
    0 references
    stochastic invariance
    0 references
    Stratonovich drift
    0 references
    stochastic control systems
    0 references
    deterministic control system with two controls
    0 references
    stochastic control system with time independent controls
    0 references
    regularity of a diffusion
    0 references
    state constraints
    0 references
    viability
    0 references
    0 references