MODELS FOR PREDICTION AND CONTROL IX DYNAMIC MODELS.
WISCONSIN UNIV MADISON DEPT OF STATISTICS
Pagination or Media Count:
In previous chapters we have considered the problem of representing practically occurring time series by linear stochastic difference equations. As a first example of the use of such models we have shown how they may be used for forecasting both non-seasonal and seasonal series. These series may also be used to represent disturbances such as arise in economics, engineering and in business applications. To control such systems we need not only an adequate disturbance model, but also a dynamic model which can take account of the various types of inertia which may be inherent in the process. The present chapter provides an introduction to models based on difference equations which give a parsimonious mathematical representation of the dynamic behaviour of physical systems. Two cases will be distinguished 1 When the input to the system is kept fixed over constant intervals of time as in the discrete control systems, and, 2 When an essentially continuous input and output are sampled at equidistant time intervals and the resulting discrete time series used to estimate the dynamic response or transfer function of the system.
- Numerical Mathematics