An Exact Estimator-Controller Solution to a Stochastic Optimal-Control Problem with Point Process Observations.
Abstract
The so-called dual control problem is the most general stochastic optimal-control problem and has been solved only under very restrictive conditions. Of special importance is the separation theorem which demonstrates that for a linear stochastic plant, quadratic costs, and linear observations in additive Gaussian noise, the optimal control law can be determined by solving separately and independently a causal stochastic-estimation problem and a deterministic control problem. In this paper, we give the exact solution to a dual control problem involving a linear stochastic plant, quadratic costs, and nonlinear, nongaussian observations. The observations are in the form of a point process in which each point has both a temporal and a spatial coordinate. The state of the stochastic plant influences the intensity of the observed time-space point process. We show that the solution to this dual control problem can be realized with a separated estimator-controller in which the estimator is nonlinear, mean-square optimal, and finite-dimensional, and the controller is linear.
Document Details
- Document Type
- Technical Report
- Publication Date
- Sep 01, 1975
- Accession Number
- ADA019555
Entities
People
- Donald L. Snyder
- Estil V. Hoversten
- Ian B. Rhodes
Organizations
- University of Washington