Title :
Stochastic dynamic programming: advanced computing constructs
Author_Institution :
Dept. of Math., Stat. & Comput. Sci., Illinois Univ., Chicago, IL, USA
Abstract :
A method for fast and efficient numerical solution of stochastic dynamic programming problems is described. The problems arise in optimal control of nonlinear, continuous-time dynamical systems, perturbed by Poisson as well as Gaussian random white noise. The numerical formulation is highly suitable for a vector multiprocessor or vectorizing supercomputer. Advanced computing techniques and hardware help alleviate Bellman´s curse of dimensionality in dynamic programming computations
Keywords :
computational complexity; dynamic programming; nonlinear control systems; numerical methods; optimal control; stochastic programming; Gaussian random white noise; Poisson noise; advanced computing constructs; continuous-time systems; curse of dimensionality; dynamical systems; efficient numerical solution; nonlinear systems; optimal control; stochastic dynamic programming; vector multiprocessor; vectorizing supercomputer; Cost function; Dynamic programming; Feedback control; Optimal control; Stochastic processes; Stochastic resonance; Stochastic systems; Stress control; Switches; White noise;
Conference_Titel :
Decision and Control, 1989., Proceedings of the 28th IEEE Conference on
Conference_Location :
Tampa, FL
DOI :
10.1109/CDC.1989.70252
