Sensitivity and parameter identifiability in linear systems
โ Scribed by G. Guardabassi; F. Romeo; R. Scattolini
- Publisher
- Elsevier Science
- Year
- 1981
- Tongue
- English
- Weight
- 537 KB
- Volume
- 312
- Category
- Article
- ISSN
- 0016-0032
No coin nor oath required. For personal study only.
โฆ Synopsis
In this paper a class of discrete linear systems affected by a time varying parameter are considered. By means of a sensitivity analysis efective parameter identifiability conditions are obtained. An application to the identification of a cross-flow heat exchanger is finally discussed.
๐ SIMILAR VOLUMES
For stability analysis of linear periodic systems with more than one degree of freedom, the Floquet method is a general and valuable, practical method. In multi-parameter periodic systems, repeated numerical integration to obtain the Floquet matrix may be a limiting factor, and e!ective sensitivity
Tbe sensitivity of performance index with respect to parameter variations in the state equation of optimally designed linear time-delay systems is discussed by introducing the concept of p-sensitivity for such systems. Two cases are considered depending on the nature of these variations. In one case
This paper is concerned with the sensitivity of a linear system-quadratic type cost functional to large parameter variations in the description of the linear system. For some class of parameter variations e and some real number p, the concept of p-sensitivity is introduced: /)-sensitivity occurs ff
We consider a controlled stochastic semilinear evolution equation with the drift depending on the unknown parameter. We show that the maximum likelihood estimator is strongly consistent for a class of bounded predictable controls.
A new measure of sensitivity with respect to initial conditmns in linear multi-input time-varying systems is defined. Remarkable amongst its properties are its nondependence on open-loop control and its invariance under a regular linear time-invanant transformation of the states. A system is shown t