An Estimation Problem of Chemical Process with Conuent Parameters: An Interval Approach
S.I. Kumkov, N. N. Krasovskiĭ, B. N. Eltsin · 2016
We apply interval analysis to a practical problem of estimation of conuent (merged) parameters of a chemical process under the following uncertainty conditions: the process measurements are noisy (corrupted), the probability characteristics of measuring bounded errors are absent, the measurement sample is fatally short, and only approximate a priori intervals of the parameters are available. Our algorithm uses a model function describing the chemical process and produces a containment set of admissible values of the parameters with an exact description of its boundaries in comparison with the known parallelotope estimation approach.