Automated synthesis of multiple analog circuits using evolutionary computation for redundancy-based fault-tolerance

Analog circuits are one of the most important parts of modern electronic systems and the failure of electronic hardware presents a critical threat to the completion of modern aircraft, spacecraft, and robot missions. Compared to digital circuits, designing fault-tolerant analog circuits is a difficult and knowledge-intensive task. A simple but powerful method for robustness is a redundancy approach to use multiple circuits instead of single one. For example, if component failures occur, other redundant components can replace the functions of broken parts and the system can still work. However, there are several research issues to make the redundant system automatically. In this paper, we used evolutionary computation to generate multiple analog circuits automatically and then we combined the solutions to generate robust outputs. Evolutionary computation is a natural way to produce multiple redundant solutions because it is a population-based search. Experimental results on the evolution of the low-pass, high-pass and band-stop filters show that the combination of multiple evolved analog circuits produces results that are more robust than those of the best single circuit.     

A comprehensive survey on functional link neural networks and an adaptive PSO–BP learning for CFLNN

This paper proposed a hybrid genetic based functional link artificial neural network (HFLANN) with simultaneous optimization of input features for the purpose of solving the problem of classification in data mining. The aim of the proposed approach is to choose an optimal subset of input features using genetic algorithm by eliminating features with little or no predictive information and increase the comprehensibility of resulting HFLANN. Using the functionally expanded of selected features, HFLANN overcomes the nonlinearity nature of problems, which is commonly encountered in single-layer neural networks. The features like simplicity of the architecture and low computational complexity of the network encourage us to use it in classification task of data mining. Further, the issue of statistical tests for comparison of algorithms on multiple datasets, which is even more essential to typical machine learning and data mining studies, has been all but ignored. In this work, we recommend a set of simple, yet safe and robust parametric and nonparametric tests for statistical comparisons of HFLANN with FLANN and RBF classifiers over multiple datasets by an extensive simulation studies.     

An Evolutionary Algorithm Approach to Optimal Ensemble Classifiers for DNA Microarray Data Analysis

In general, the analysis of microarray data requires two steps: feature selection and classification. From a variety of feature selection methods and classifiers, it is difficult to find optimal ensembles composed of any feature-classifier pairs. This paper proposes a novel method based on the evolutionary algorithm (EA) to form sophisticated ensembles of features and classifiers that can be used to obtain high classification performance. In spite of the exponential number of possible ensembles of individual feature-classifier pairs, an EA can produce the best ensemble in a reasonable amount of time. The chromosome is encoded with real values to decide the weight for each feature-classifier pair in an ensemble. Experimental results with two well-known microarray datasets in terms of time and classification rate indicate that the proposed method produces ensembles that are superior to individual classifiers, as well as other ensembles optimized by random and greedy strategies.     

Flexural and shear behaviors of welded H-beam fabricated with hot-rolled steel coils

HR (Hot-Rolled steel coil) plates are produced by straightening hot-rolled steel coil with a primary objective to substitute conventional thick plates, thereby avoiding a number of difficulties encountered in the production and shipment of thick plates and relieving supply problems. In order to evaluate the applicability of HR plates to building structures under KBC 2009 steel design provision, a total of four full-scale welded built-up H-beam with simple supports are tested together with associated key material testing. The full-scale beam specimens are respectively classified into two categories; the flexure-dominated beam and the shear-dominated beam; to evaluate both the bending and shear strength separately. In the material test the HR plates satisfy both the tensile mechanical properties and the CVN toughness which are required in the current material specification. Both the flexure-dominated and the shear-dominated beams exhibit sufficient capacities that exceed the nominal strengths specified in the KBC 2009, thus confirming the applicability of the HR plate.     

Evolutionary ensemble of diverse artificial neural networks using speciation

Recently, many researchers have designed neural network architectures with evolutionary algorithms but most of them have used only the fittest solution of the last generation. To better exploit information, an ensemble of individuals is a more promising choice because information that is derived from combining a set of classifiers might produce higher accuracy than merely using the information from the best classifier among them. One of the major factors for optimum accuracy is the diversity of the classifier set. In this paper, we present a method of generating diverse evolutionary neural networks through fitness sharing and then combining these networks by the behavior knowledge space method. Fitness sharing that shares resources if the distance between the individuals is smaller than the sharing radius is a representative speciation method, which produces diverse results than standard evolutionary algorithms that converge to only one solution. Especially, the proposed method calculates the distance between the individuals using average output, Pearson correlation and modified Kullback–Leibler entropy to enhance fitness sharing performance. In experiments with Australian credit card assessment, breast cancer, and diabetes in the UCI database, the proposed method performed better than not only the non-speciation method but also better than previously published methods.     

Near-Infrared Imaging of Steroid Hormone Activities Using Bright BRET Templates

Bioluminescence (BL) is an excellent optical readout for bioassays and molecular imaging. Herein, we accomplished new near infrared bioluminescence resonance energy transfer (NIR-BRET) templates for monitoring molecular events in cells with higher sensitivity. We first identified the best resonance energy donor for the NIR-BRET templates through the characterization of many coelenterazine (CTZ)–marine luciferase combinations. As a result, we found that NLuc–DBlueC and ALuc47–nCTZ combinations showed luminescence in the blue emission wavelength with excellent BL intensity and stability, for example, the NLuc–DBlueC and ALuc47–nCTZ combinations were 17-fold and 22-fold brighter than their second highest combinations, respectively, and were stably bright in living mammalian cells for at least 10 min. To harness the excellent BL properties to the NIR-BRET systems, NLuc and ALuc47 were genetically fused to fluorescent proteins (FPs), allowing large “blue-to-red” shifts, such as LSSmChe, LSSmKate2, and LSSmNep (where LSS means Large Stokes Shift). The excellent LSSmNep–NLuc combination showed approximately 170 nm large resonance energy shift from blue to red. The established templates were further utilized in the development of new NIR-BRET systems for imaging steroid hormone activities by sandwiching the ligand-binding domain of a nuclear receptor (NR-LBD) between the luciferase and the FP of the template. The NIR-BRET systems showed a specific luminescence signal upon exposure to steroid hormones, such as androgen, estrogen, and cortisol. The present NIR-BRET templates are important additions for utilizing their advantageous imaging of various molecular events with high efficiency and brightness in physiological samples.     

Mobile context inference using two-layered Bayesian networks for smartphones

Recently, mobile context inference becomes an important issue. Bayesian probabilistic model is one of the most popular probabilistic approaches for context inference. It efficiently represents and exploits the conditional independence of propositions. However, there are some limitations for probabilistic context inference in mobile devices. Mobile devices relatively lacks of sufficient memory. In this paper, we present a novel method for efficient Bayesian inference on a mobile phone. In order to overcome the constraints of the mobile environment, the method uses two-layered Bayesian networks with tree structure. In contrast to the conventional techniques, this method attempts to use probabilistic models with fixed tree structures and intermediate nodes. It can reduce the inference time by eliminating junction tree creation. To evaluate the performance of this method, an experiment is conducted with data collected over a month. The result shows the efficiency and effectiveness of the proposed method.     

Numerical evaluation of the retrofit effectiveness for GFRP retrofitted concrete slab subjected to blast pressure

For retrofitting structures against blast loads, sufficient ductility and strength should be provided by using high-performance materials such as fiber reinforced polymer (FRP) composites. The effectiveness of retrofit materials needs to be precisely evaluated for the retrofitting design based on the dynamic material responses under blast loads. In this study, refined FEM analysis with high-strain rate dependent material model and debonding failure model is conducted for evaluating the FRP retrofitting effectiveness. The structural behavior of reinforced concrete (RC) slab retrofitted with glass fiber reinforced polymer (GFRP) under blast pressure is simulated and the analysis results are verified with the previous experimental results.     

Semantic management of multiple contexts in a pervasive computing framework

Mobile devices can perceive greater details of user states with the increasing integration of mobile sensors into a pervasive computing framework, yet they consume large amounts of batteries and computational resources. This paper proposes a semantic management method which efficiently integrates multiple contexts into the mobile system by analyzing the semantic hierarchy and temporal relations. The proposed method semantically decides the recognition order of the contexts and identifies each context using a corresponding dynamic Bayesian network (DBN). To sort out the contexts, we designed a semantic network using a knowledge-driven approach, whereas DBNs are constructed with a data-driven approach. The proposed method was validated on a pervasive computing framework, which included multiple mobile sensors (such as motion sensors, data-gloves, and bio-signal sensors). Experimental results showed that the semantic management of multiple contexts dramatically reduced the recognition cost.     

DE+RBFNs based classification: A special attention to removal of inconsistency and irrelevant features

A novel approach for the classification of both balanced and imbalanced dataset is developed in this paper by integrating the best attributes of radial basis function networks and differential evolution. In addition, a special attention is given to handle the problem of inconsistency and removal of irrelevant features. Removing data inconsistency and inputting optimal and relevant set of features to a radial basis function network may greatly enhance the network efficiency (in terms of accuracy), at the same time compact its size. We use Bayesian statistics for making the dataset consistent, information gain theory (a kind of filter approach) for reducing the features, and differential evolution for tuning center, spread and bias of radial basis function networks. The proposed approach is validated with a few benchmarked highly skewed and balanced dataset retrieved from University of California, Irvine (UCI) repository. Our experimental result demonstrates promising classification accuracy, when data inconsistency and feature selection are considered to design this classifier.