This paper presents a controller design for shaping conditional output probability density functions (pdf) for non-Gaussian dynamic stochastic systems whose coefficients are random and represented by their known pdfs. The moment-generating function is applied to all the pdfs, leading to a simple mathematical relationship amongst all the transferred conditional pdfs of the system output and random parameters. A new performance function is introduced and its minimisation is performed so as to design an optimal control input that makes the shape of the conditional output pdf follow a target distribution. An example is included to illustrate the use of the algorithm. Copyright © 2011 Inderscience Enterprises Ltd.
|Number of pages
|International Journal of Systems, Control and Communications
|Published - Mar 2011
- Dynamic stochastic systems
- Gradient-based optimisation
- Probability density functions
- The moment-generating function