Observation control for discrete-continuous stochastic systems
The author considers a partially observable process \(\{x(t),y(t),z(t)\}\) in \(\mathbb{R}^n\times \mathbb{R}^m\times \mathbb{R}^k\) satisfying the system of equations with controllable observations \[ \begin{aligned} dx(t)&= dA(t)x(t-)+d\xi_t, \quad x(0-)=x_0,\\ dy(t)&=\alpha\big(t,u(t),z(t)\big)x(t)v(t) dt+d\eta_t,\quad y(0)=y_0,\\ dz(t)&=f\big(t,u(t),z(t)\big)x(t)v(t) dt +G\big(t,u(t),z(t)\big)v(t) dt,\quad z(0)=z_0.\end{aligned} \] Here \(x\) is an unobservable process; \(y\) and \(z\) are observable control processes satisfying certain constrainst; \(A=A(t)\) is a matrix-valued function of the form \[ A(t)=\int_0^ta(s) ds+\sum_{\tau_i\leq t}\Delta A(\tau_i) \] with continuous \(a\) and finite number of jumps \(\Delta A(\tau_i)\); the initial state \(x_0\) (a Gaussian random variable), and the driving processes \(\xi\) and \(\eta\) (Gaussian martingales) are assumed to be mutually independent. The problem of the best estimation of the process \(x\) with discrete and continuous observations is reduced to a deterministic one with impulse or generalized control. An approach based on the discontinuous time transformation is presented and used to reduce the problem with impulsive control to one with bounded controls. The existence and description of an optimal generalized observation process are also discussed. The main novelty of the paper is that discontinuous observable processes are considered (and, therefore, the known continuous-case results cannot be applied because of the possible coincidence of jumps and observation moments).
- On the construction of nearly optimal strategies for a general problem of control of partially observed diffusions
- Controllable Markov jump processes. I: Optimum filtering based on complex observations
- An Approach to Discrete-Time Stochastic Control Problems under Partial Observation
- Stochastic continuous control of partially observed systems via impulse control problems
- Existence of optimal controls for partially observed jump processes
- Controllable Markov jump processes. I: Optimum filtering based on complex observations
- Analysis and computation of an optimality equation arising in an impulse control problem with discrete and costly observations
- Hybrid method for a general optimal sensor scheduling problem in discrete time
- Control of dynamical systems with discrete and uncertain observations
- Robust design of sensor fusion problem in discrete time
- Impulsive control with impulsive actions of two types
- scientific article; zbMATH DE number 4181271 (Why is no real title available?)
- scientific article; zbMATH DE number 4007380 (Why is no real title available?)
- scientific article; zbMATH DE number 4064863 (Why is no real title available?)
- Discrete-scanning observation problem for stochastic non-linear discrete-time distributed parameter systems
- scientific article; zbMATH DE number 484192 (Why is no real title available?)
This page was built for publication: Observation control for discrete-continuous stochastic systems
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4507157)
