Evolution of neural networks for classification and regression

Although Artificial Neural Networks (ANNs) are importantdata mining techniques, the search for the optimal ANN is a challenging task: the ANN should learn the input–output mapping without overfitting the data and training algorithms may get trapped in local minima. The use of Evolutionary Computation (EC) is a promising alternative for ANN optimization. This work presents two hybrid EC/ANN algorithms: the first evolves neural topologies while the latter performs simultaneous optimization of architectures and weights. Sixteen real-world tasks were used to test these strategies. Competitive results were achieved when compared with a heuristic model selection and other Data Mining algorithms.     

Binding of an arm repeat protein to the kinase domain of the S-locus receptor kinase

Screening of a yeast two-hybrid library for proteins that interact with the kinase domain of an S-locus receptor kinase (SRK) resulted in the isolation of a plant protein called ARC1 (Arm Repeat Containing). This interaction was mediated by the C-terminal region of ARC1 in which five arm repeat units were identified. Using the yeast two-hybrid system and in vitro binding assays, ARC1 was found to interact specifically with the kinase domains from SRK-910 and SRK-A14 but failed to interact with kinase domains from two different Arabidopsis receptor-like kinases. In addition, treatment with a protein phosphatase or the use of a kinase-inactive mutant reduced or abolished the binding of ARC1 to the SRK-910 kinase domain, indicating that the interaction was phosphorylation dependent. Lastly, RNA blot analysis revealed that the expression of ARC1 is restricted to the stigma, the site of the self-incompatibility response.     

13C NMR study of presence of uron structures in amino adhesives and relation with wood‐based panels performance

The chemical structure of amino adhesives produced by the strongly acid process was investigated by ¹³C Nuclear Magnetic Resonance spectroscopy. This technique allowed the identification of functional groups and its quantitative determination. Concentration of cyclic polymeric structures (urons) was shown to be related with adhesive performance and with particleboard physico‐mechanical properties and formaldehyde content. Higher urons concentration presented lower viscosity and reactivity. Particleboards produced with resins with lower urons concentration presented lower formaldehyde content, but also lower internal bond strength. Wood‐based panels produced fulfilled E1 class requirements for formaldehyde emissions, indicating that strongly acid process is an alternative to the conventional alkaline–acid process. © 2013 Wiley Periodicals, Inc. J. Appl. Polym. Sci. 130: 4500–4507, 2013     

Online Identification of Horizontal Two-Phase Flow Regimes Through Gabor Transform and Neural Network Processing

The fundamental objective of this work is the development of a specialist system capable of diagnosing different configurations of horizontal two-phase flow regimes. It is important to emphasize that this knowledge is of capital importance to the efficient operation of facilities for the manipulation and transportation of multiphase fluids, and it represents one of the most important challenges in petrochemical and thermonuclear industries today. The working principle of the proposed methodology is based on the signals acquired by a rapid response pressure gradient sensor and their decomposition into Gabor coefficients, followed by processing through a previously trained artificial neural network. The implementation is accomplished in such a way that the diagnosis operation is performed online, from the acquisition of the signal to its post-processing. An experimental campaign was conducted at the facilities of the Thermal and Fluids Engineering Laboratory (NETeF) at the University of São Paulo in order to validate the proposed methodology. Experimental pressure gradient signals were obtained for all main horizontal air-water flow regimes (stratified smooth, stratified wavy, intermittent, annular, and bubbly) produced in a 12 m long test section with an internal diameter of 30 mm. Results show that the percentage of correct flow regime diagnosis in steady-state conditions is practically 100%, provided the detection level is adequately set.     

(Virtual) Agents for running electricity markets

This paper describes a multi-agent based simulation (MABS) framework to construct an artificial electric power market populated with learning agents. The artificial market, named TEMMAS (The Electricity Market Multi-Agent Simulator), explores the integration of two design constructs: (i) the specification of the environmental physical market properties and (ii) the specification of the decision-making (deliberative) and reactive agents. TEMMAS is materialized in an experimental setup involving distinct power generator companies that operate in the market and search for the trading strategies that best exploit their generating units’ resources. The experimental results show a coherent market behavior that emerges from the overall simulated environment.     

Frequency domain analysis of the mirror-modified filtered-x least mean squares algorithm with low ambient noise

The mirror-modified filtered-x least mean squares (MMFxLMS ) algorithm is a variation of the FxLMS algorithm with online secondary path modeling that cannot diverge due to secondary path modeling errors. However, problems may occur when the ambient noise is not limited due to insufficient modeling power. This work shows that under a frequency domain analysis without ambient noise, the MMFxLMS algorithm is always stable, and expressions for the maximum residual noise level at any given time are obtained. It is also shown that, under the same context, convergence to the minimum residual noise is guaranteed. Still, convergence can be much slower for high secondary path modeling errors than that of the LMS or MFxLMS algorithms. Simulations confirm these results.     

Giant electric-field-induced strains in lead-free ceramics for actuator applications – status and perspective

In response to the current environmental regulations against the use of lead in daily electronic devices, a number of investigations have been performed worldwide in search for alternative piezoelectric ceramics that can replace the market-dominating lead-based ones, representatively Pb(Zr _x Ti_1-x )O₃ (PZT)-based solid solutions. Selected systems of potential importance such as chemically modified and/or crystallographically textured (K, Na)NbO₃ and (Bi_1/2Na_1/2)TiO₃-based solid solutions have been developed. Nevertheless, only few achievements have so far been introduced to the marketplace. A recent discovery has greatly extended our tool box for material design by furnishing (Bi_1/2Na_1/2)TiO₃-based ceramics with a reversible phase transition between an ergodic relaxor state and a ferroelectric with the application of electric field. This paired the piezoelectric effect with a strain-generating phase transition and extended opportunities for actuator applications in a completely new manner. In this contribution, we will present the status and perspectives of this new class of actuator ceramics, aiming at covering a wide spectrum of topics, i.e., from fundamentals to practice.     

Spiral periodic structure inside chaotic region in parameter‐space of a Chua circuit

In this letter we investigate, via numerical simulations, the parameter‐space of the set of autonomous first‐order differential equations of a Chua circuit. We show that this parameter‐space presents self‐organized periodic structures immersed in a chaotic region, forming a single spiral structure that coils up around a focal point. Additionally, bifurcation diagrams are used to show that those periodic structures also organize themselves in period‐adding cascades, along specific directions that point towards this same focal point. Copyright © 2010 John Wiley & Sons, Ltd.