Todos gentilmente convidados para o 30o SEMINÁRIO DE CONTROLE LADES-GIMSCOP

Palestrante: Professor Miguel Bagajewicz

Link para a apresentação: meet.google.com/ffw-goue-hea

Abstract: “Back-off control relies on the use of an expected dynamic operating region (EDOR), whose shape can be accommodated by designing a good control structure. The shape and orientation of the EDOR notwithstanding, the closeness of this region to the optimal operating point is governed by the precision and bias-free nature of the data used. In turn, precision of data is a property of different filters; maximum bias is also a feature that some filters are capable of control. The closeness of the EDOR to the constraints is related to cost or in general, to profit. While precision is well defined in the literature, in practice, accuracy has many parents. I use the definition that accuracy is the sum of maximum bias + precision.  The calculation of the maximum bias, is in turn a matter of further definition, depending on the tool used to as an estimator. In the context of techniques like data reconciliation, we consider some means of detecting bias and the calculation of its maximum. However, the maximum bias is often too large, so one should use maximum EXPECTED bias. This leads to the definition of Stochastic Accuracy. Knowing the stochastic accuracy helps knowing the size and shape of the EDOR”.

Minibio: Miguel Bagajewicz é Professor Visitante da UFRJ (EPQB/EQ) e Professor Emérito (previamente “Sam Wilson Professor & Samuel Roberts Noble Foundation Presidential Professor”) da Universidade de Oklahoma, EUA.