Towards online computation of information state controllers
Nonlinear \(H_{\infty}\) control theory is based on two partial differential equations. One is the first-order evolution equation called the information state equation, which plays the role of a generalized state estimator and constitutes the dynamics of the controller. It must be implemented online. The second partial differential equation is defined on the space of possible information states and determines the output law for the controller. It can be implemented offline. This paper discusses the information state partial differential equation only. Two threads of research relating to online computation are described: (i) numerical techniques using max-plus expansions, which are well matched to the mathematical structure of information state dynamics, and (ii) the ``cheap sensor problem, which has dramatically less complexity than the regular case. For the entire collection see [Zbl 0966.00045].
- Optimal on-line control under imperfect information
- Effective information for offline stochastic feedback and optimal control of dynamic systems
- Effective information in offline stochastic feedback and optimal control of dynamic systems: results and applications
- Control and estimation under information constraints: toward a unified theory of control, computation and communications
- State-dependent sampling for online control
- Online State Estimation for Time-Varying Systems
- scientific article; zbMATH DE number 4160585
- Information states in stochastic control and filtering: a Lie algebraic theoretic approach
- Near-optimal online control of dynamic discrete-event systems
This page was built for publication: Towards online computation of information state controllers
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2758356)
