Optimization of neural classifiers based on bayesian decision boundaries and idle neurons pruning

In this article we describe a feature extraction algorithm for pattern classification based on Bayesian Decision Boundaries and Pruning techniques. The proposed method is capable of optimizing MLP neural classifiers by retaining those neurons in the hidden layer that realy contribute to correct classification. Also in this article we proposed a method which defines a plausible number of neurons in the hidden layer based on the stem-and-leaf graphics of training samples. Experimental investigation reveals the efficiency of the proposed method. © 2002 IEEE.

Roll waves evolution in high gradient channels to a non-Newtonian rheology

The criteria for the occurrence of roll wave phenomenon in the supercritical and turbulent Newtonian and non-Newtonian flows from the engineering point of view was analyzed. Imposing a constant discharge at the upstream of the canal and superposing a small perturbation, it was observed that roll waves can be developed more easily for small wave numbers and for high cohesions. Moreover, from the mathematical model used, it was demonstrated that the numerical viscosity was 10 times the physical viscosity.

Water waves generated by local disturbance: Serre's model validity

Water waves generated by landslides were long menace in certain localities and the study of this phenomenon were carried out at an accelerated rate in the last decades. Nevertheless, the phase of wave creation was found to be very complex. As such, a numerical model based on Boussinesq equations was used to describe water waves generated by local disturbance. This numerical model takes in account the vertical acceleration of the particles and considers higher orders derivate terms previously neglected by Boussinesq, so that in the generation zone, this model can support high relative amplitude of waves.

Controlling collapse in Bose-Einstein condensates by temporal modulation of the scattering length

A study was conducted on the dynamics of 2D and 3D Bose-Einstein condensates in the case when the scattering length in the Gross-Pitaevskii (GP) equation which contains constant (dc) and time-variable (ac) parts. Using the variational approximation (VA), simulating the GP equation directly, and applying the averaging procedure to the GP equation without the use of the VA, it was demonstrated that the ac component of the nonlinearity makes it possible to maintain the condensate in a stable self-confined state without external traps.

Araçatuba virus: A vaccinialike virus associated with infection in humans and cattle

We describe a vaccinialike virus, Araçatuba virus, associated with a cowpoxlike outbreak in a dairy herd and a related case of human infection. Diagnosis was based on virus growth characteristics, electron microscopy, and molecular biology techniques. Molecular characterization of the virus was done by using polymerase chain reaction amplification, cloning, and DNA sequencing of conserved orthopoxvirus genes such as the vaccinia growth factor (VGF), thymidine kinase (TK), and hemagglutinin. We used VGF-homologous and TK gene nucleotide sequences to construct a phylogenetic tree for comparison with other poxviruses. Gene sequences showed 99% homology with vaccinia virus genes and were clustered together with the isolated virus in the phylogenetic tree. Araçatuba virus is very similar to Cantagalo virus, showing the same signature deletion in the gene. Araçatuba virus could be a novel vaccinialike virus or could represent the spread of Cantagalo virus.

Adaptive control charts for attributes

We develop a general model for adaptive c, np, u and p control charts in which one, two or three design parameters (sample size, sampling interval and control limit width) switch between two values, according to the most recent process information. For a given in-control average sampling rate and a given false alarm rate, the adaptive chart detects changes in the process much faster than a chart with fixed parameters. Moreover, this study also offers general guidance on how to choose an effective design.

Scaling predictions for radii of weakly bound triatomic molecules

The mean-square radii of the triatomic molecules 4He 3, 4He 2- 6Li, 4He 2- 7Li, and 4He 2- 23Na were calculated using a renormalized three-body model with a pairwise Dirac-δ interaction, having as physical inputs only the values of the binding energies of the diatomic and triatomic molecules. Molecular three-body systems with bound subsystems were considered. The resultant data were analyzed in detail.

A Short Note on a Nonlinear System vibrations under Two Non-Ideal Excitations

This paper describes a nonlinear phenomenon in the dynamical behavior of a nonlinear system under two non-ideal excitations: the self-synchronization of unbalanced direct current motors. The considered model is taken as a Duffing system that is excited by two unbalanced direct current motors with limited power supplies. The results obtained by using numerical simulations are discussed in details.

Array of Bose-Einstein condensates under time-periodic Feshbach-resonance management

The investigation of the dynamics of a discrete soliton in an array of Bose-Einstein condensates under the action of a periodically time-modulated atomic scattering length [Feshbach-resonance management (FRM)] was discussed. The slow and rapid modulations, in comparison with the tunneling frequency were considered. An averaged equation, which was a generalized discrete nonlinear Schrödinger equation, including higher-order effective nonlinearities and intersite nonlinear interactions was derived in the case of the rapid modulation. It was demonstrated that the modulations of sufficient strength results in splitting of the soliton by direct simulations.

A FEM approach to the equations of magneto-aerodynamics

In this work, a Finite Element Method treatment is outlined for the equations of Magnetoaerodynamics. In order to provide a good basis for numerical treatment of Magneto-aerodynamics, a full version of the complete equations is presented and FEM contribution matrices are deduced, as well as further terms of stabilization for the compressible flow case.

Tuning of fuzzy inference systems through unconstrained optimization techniques

This paper presents a new methodology for the adjustment of fuzzy inference systems. A novel approach, which uses unconstrained optimization techniques, is developed in order to adjust the free parameters of the fuzzy inference system, such as its intrinsic parameters of the membership function and the weights of the inference rules. This methodology is interesting, not only for the results presented and obtained through computer simulations, but also for its generality concerning to the kind of fuzzy inference system used. Therefore, this methodology is expandable either to the Mandani architecture or also to that suggested by Takagi-Sugeno. The validation of the presented methodology is accomplished through an estimation of time series. More specifically, the Mackey-Glass chaotic time series estimation is used for the validation of the proposed methodology.

On Nonlinear Dynamics and Control of A Particular Portal Frame Foundation Model, Excited by a Non-Ideal Motor

The nonlinear dynamic response and a nonlinear control method of a particular portal frame foundation for an unbalanced rotating machine with limited power (non-ideal motor) are examined. Numerical simulations are performed for a set of control parameters (depending on the voltage of the motor) related to the static and dynamic characteristics of the motor. The interaction of the structure with the excitation source may lead to the occurrence of interesting phenomena during the forward passage through the several resonance states of the systems. A mathematical model having two degrees of freedom simplifies the non-ideal system. The study of controlling steady-state vibrations of the non-ideal system is based on the saturation phenomenon due to internal resonance.

Behavioural Analysis of a Nonlinear Mechanical System Using Transient Trajectories

This work aims at a better comprehension of the features of the solution surface of a dynamical system presenting a numerical procedure based on transient trajectories. For a given set of initial conditions an analysis is made, similar to that of a return map, looking for the new configuration of this set in the first Poincaré sections. The mentioned set of I.C. will result in a curve that can be fitted by a polynomial, i.e. an analytical expression that will be called initial function in the undamped case and transient function in the damped situation. Thus, it is possible to identify using analytical methods the main stable regions of the phase portrait without a long computational time, making easier a global comprehension of the nonlinear dynamics and the corresponding stability analysis of its solutions. This strategy allows foreseeing the dynamic behavior of the system close to the region of fundamental resonance, providing a better visualization of the structure of its phase portrait. The application chosen to present this methodology is a mechanical pendulum driven through a crankshaft that moves horizontally its suspension point.

Mean-field model of jet formation in a collapsing Bose-Einstein condensate

We perform a systematic numerical study, based on the time-dependent Gross-Pitaevskii equation, of jet formation in collapsing and exploding Bose-Einstein condensates as in the experiment by Donley et al (2001 Nature 412 295). In the actual experiment, via a Feshbach resonance, the scattering length of atomic interaction was suddenly changed from positive to negative on a pre-formed condensate. Consequently, the condensate collapsed and ejected atoms via explosion. On disruption of the collapse by suddenly changing the scattering length to zero, a radial jet of atoms was formed in the experiment. We present a satisfactory account of jet formation under the experimental conditions and also make predictions beyond experimental conditions which can be verified in future experiments.

Pressure field in the cathode-anode region of a high-power klystron amplifier

The pressure field of a high-power klystron amplifier in the cathode and anode region was investigated. The investigation was performed using a 1.3 GHz, 100 A and 240 kV high-power klystron with five reentrant coaxial cavities, assembled in cylindrical drift tube 1.2 m long. The diffusion equation in mathematical model was also solved by using a 3-D finite element method code, in order to obtain pressure profile in region of interest. The results show that density profile of molecules between cathode-anode region was determined, where cathode pressure is approximately 10% higher than anode pressure.