Treatment of Uncertainties with Function Signals obtained by configurations of Paraconsistent Artificial Neural Cells Arnaldo de Carvalho Jr, Hyghor M. Côrtes, Clovis M. da Cruz, Dorotéa Vilanova Garcia, Maurício C. Mario, João Inácio da Silva Filho

Abstract

 Treatments of uncertainties show better results when non-classical logic is used. The Paraconsistent Annotated Logic with annotation of two values (PAL2v) is one type of non-classical logics that differs from classical logic (binary) by allowing the processing of signals of contradictory information. A Paraconsistent Artificial Neural Cell (PANcel) is an algorithm that uses PAL2v to simulate neuron behavior. Applying the degrees of favorable evidence (µ) and unfavorable evidence  (l) on its two inputs, it presents at its outputs the degree of resulting analysis evidence (µ_E) or real analysis evidence (µ_RE) and the resulting evidence interval (j_E). A Paraconsistent Artificial Neural Cell of Learning (LPANcel) and its variations is a type of PANcel that uses the output as a feedback to the input of the unfavorable degree of evidence in time, it can learn any real value within a closed range (interval values [0,1]). This cell can be used in signal analysis and treatment. In this work, the results of simulations of LPANcel are presented and its variations as integrator and differentiator, what may be a viable alternative to reproducing signals of functions used in automation control systems.