Identification of Parametric Models: From Experimental DataSpringer, 1997. gada 14. janv. - 413 lappuses The identification of parametric models from experimental data is a fundamental activity among researchers and engineers in pure and applied sciences. This work addresses the topic by examining, among others, the following areas: • choice of an appropriate model structure which allows the estimation of all parameters; • choice of a quality criterion for rating models; • incorporation of prior knowledge and objectives and guarding against possible outliers; • optimization of the selected criterion and simple yet exact evaluation of characteristics; • evaluation of uncertainty in estimated parameters; • design of experimental conditions for the collection of the most pertinent information given prior constraints and objectives. Identification of Parametric Models deals with these questions in a straightforward style while providing a global vision of the methodology. Suitable for engineers and researchers who practise mathematical modelling from experimental data, graduate students who wish to become acquainted with the field, this text will also be a valuable resource for specialists in the field. |
No grāmatas satura
1.–3. rezultāts no 30.
... optimization problems often have the following characteristics : the number of parameters to be optimized is small ... global optimizers , should be taken into account . The optimization may , on the other hand , be centred on the cost ...
... global optimization methods . The initial problem is generally decomposed into a sequence of more elementary problems . For instance , a multivariable optimization problem is considered as a sequence of one - dimensional optimizations ...
... Global optimization Global optimization techniques aim to find the best possible value for the cost and the associated optimizer ( s ) p , such that for any feasible p j ( p ) ≥ j ( p ) = } . Insofar as they succeed , they bypass the ...