Time Series Analysis: Forecasting and Control, Fourth Edition

PART FOUR

Design of Discrete Control Schemes

In earlier chapters we studied the modeling of discrete time series and dynamic systems. We saw how once adequate models have been obtained, they may be put to use to yield forecasts of time series, to characterize the transfer function of a dynamic system, and to represent the interrelationships among several time series of a multivariate dynamic system. However, the models and the methods for their manipulation are of much wider importance than even these applications indicated. The ideas we have outlined are of importance in the analysis of a wide class of stochastic–dynamic systems occurring, for example, in economics, engineering, commerce, and in organizational studies.

It is obviously impossible to illustrate every application. Rather, it is hoped that the theory and examples of this book may help the reader to adapt the general methods to their own particular problems. In doing this, the dynamic and stochastic models we have discussed will often act as building blocks that can be linked together to represent the particular system under study. Also, techniques of identification, estimation, and diagnostic checking, similar to those we have illustrated, will be needed to establish the model. Finally, recursive calculations and the ideas considered under the general heading of forecasting will have wider application in working out the consequences of a model once it has been fitted.

We shall conclude this book by illustrating these ...