INCLUSION - PROBABILITY - OPTIMAL CONTROL,
Abstract
A multi-dimensional stochastic dynamical system is considered. The system can be nonlinear and time-varying. It is assumed that the state of this system can be measured noisefree at all times. An optimal control policy is formulated leading to the maximization of the probability of constraining the trajectory of the system to a certain subregion of the state-space over a finite period of time. On the basis of dynamic programming an optimization equation is derived. This equation is a quasilinear second order parabolic partial differential equation. The equation is solved on a digital computer. The numerical technique appears to be highly efficient in speed and memory requirements. Examples are included for first and second order dynamical systems. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- May 25, 1967
- Accession Number
- AD0655387
Entities
People
- Leo J. Van Mellaert
Organizations
- New York University Tandon School of Engineering