Abstract The multi-layer perceptron has proved its efficiencies in several fields as pattern and voice recognition. Unfortunately, the classical training for MLP suffers from a poor generalization. In this respect, we have proposed a new multiobjective training model with constraints which satisfies two objectives. The first one is the learning objective: minimizing the perceptron error and the second is the complexity objective: optimizing number... The multi-layer perceptron has proved its efficiencies in several fields as pattern and voice recognition. Unfortunately, the classical training for MLP suffers from a poor generalization. In this respect, we have proposed a new multiobjective training model with constraints which satisfies two objectives. The first one is the learning objective: minimizing the perceptron error and the second is the complexity objective: optimizing number of weights and neurons. The proposed model will provide a balance between the multi-layer perceptron learning and the complexity to get a good generalization. Our model has been solved using an evolutionary approach called the Non-Dominated Sorting Genetic Algorithm (NSGA II). This approach has led to a good representation of the Pareto set for the MLP network, from which an improved generalization performance model is selected. Read More Read Less

Abstract The article is devoted to the approach to the construction of the inverse of the digital filter, which restores the original signal from the signal, distorted with the passage of the linear system. The procedure for construction provides for the use of information about the “statistical structure” of the processed signal and the characteristics of the distorting device. The efficiency of the approach... The article is devoted to the approach to the construction of the inverse of the digital filter, which restores the original signal from the signal, distorted with the passage of the linear system. The procedure for construction provides for the use of information about the “statistical structure” of the processed signal and the characteristics of the distorting device. The efficiency of the approach is illustrated by the example of the signal patterns with different statistical properties and distorting device, widely used in measurement systems. The expressions for the calculation of the coefficients are given circuit simulation and calculation results. Read More Read Less

Abstract Wind measurement is important for estimating wind energy potential but relatively cost-intensive and often conducted at a narrow height near or not far from the ground at the typical range of most wind hub heights in current wind turbine technology is 30-50 m (or even higher). Extrapolation on wind data thus becomes necessary to estimate wind speed at a high altitude. Doing so... Wind measurement is important for estimating wind energy potential but relatively cost-intensive and often conducted at a narrow height near or not far from the ground at the typical range of most wind hub heights in current wind turbine technology is 30-50 m (or even higher). Extrapolation on wind data thus becomes necessary to estimate wind speed at a high altitude. Doing so requires the essential understanding of wind shear characteristics representative to a location or a region. The analysis is carried out from the profile of meteorological data collected from 50m tower at Sathyabama University during the year 2010-2014. The tower is located near the coastal region in Chennai. The tower is equipped with instruments to measure several meteorological variables. For wind speed and direction, they are routinely measured at different heights, which are considered well suitable for wind shear characterization. In this work, the characteristics of wind shear exponent at the tower were investigated and discussed, with emphasis on temporal (diurnal and monthly) variation and occurrence distribution. Read More Read Less