Design of highly effective multilayer feedforward neural network by using genetic algorithm

This paper presents a highly effective and precise neural network method for choosing the activation functions (AFs) and tuning the learning parameters (LPs) of a multilayer feedforward neural network by using a genetic algorithm (GA). The performance of the neural network mainly depends on the learning algorithms and the network structure. The backpropagation learning algorithm is used for tuning the network connection weights, and the LPs are obtained by the GA to provide both fast and reliable learning. Also, the AFs of each neuron in the network are automatically chosen by a GA. The present study consists of 10 different functions to accomplish a better convergence of the desired input–output mapping. Test studies are performed to solve a set of two-dimensional regression problems for the proposed genetic-based neural network (GNN) and conventional neural network having sigmoid AFs and constant learning parameters. The proposed GNN has also been tested by applying it to three real problems in the fields of environment, medicine, and economics. Obtained results prove that the proposed GNN is more effective and reliable when compared with the classical neural network structure.     

A normal wiggly hesitant fuzzy linguistic projection-based multiattributive border approximation area comparison method

As a useful information representation tool, hesitant fuzzy linguistic term set (HFLTS) allows decision makers (DMs) to express their cognitive preferences in terms of several ordered and continuous linguistic terms. Considering the fact that much valuable information related to the cognitive behavior of DMs is hidden in the original evaluation information, this paper studies how to comprehensively mine uncertain information from original hesitant fuzzy linguistic evaluation information given by DMs. To address this objective, we present a new representation tool, normal wiggly hesitant fuzzy linguistic term set (NWHFLTS), which not only retains the original evaluation information, but also delivers and quantifies potential uncertain information, and can also help DMs express their evaluation information in a more complete manner. First, we develop the basic operations, score function, and comparison rule of NWHFLTS based on linguistic scale functions (LSFs), and propose the projection measure, the normal projection measure, and the normalized projection-based distance measure to describe the degree of deviation between two NWHFLTSs. Furthermore, for the case when the attribute weight is completely unknown, we combine the multiattributive border approximation area comparison (MABAC) method and develop a new method called as normal wiggly hesitant fuzzy linguistic projection-based MABAC to solve the multiattribute decision-making problems where attribute values are expressed in the form of NWHFLTS. Finally, through a practical example of marine ecological security situation, the specific calculation steps of this method are exemplified, the feasibility and advancement of the proposed method are demonstrated via a comprehensive comparative study.     

Auto-regressive modeling with dynamic weighted canonical correlation analysis

Multivariate statistical analytical methods, such as principal component analysis (PCA) and canonical correlation analysis (CCA), are widely used to extract useful information from data collected in modern industrial processes. Their dynamic extensions are also designed intensively in the literature to exploit dynamic variations in the processes. However, none of these algorithms consider the inevitably strong dependence in adjacent samples caused by high sampling frequencies or relatively slow continuous process changes, and the dependence may lead to unnecessarily large time lags and sub-optimal prediction performance. In this paper, a dynamic weighted CCA (DWCCA) algorithm is proposed to address this issue with the aid of basis functions. DWCCA constructs its latent space by maximizing the correlation between current data and a weighted representation of past data. To remove the negative effect caused by strongly collinear successive samples, it is assumed that the weights of past data rely only on a limited number of basis functions, and in this work, polynomial basis, Fourier basis, and Gaussian basis functions are considered and compared. Comprehensive modeling scheme is also designed for DWCCA with subsequent PCA decomposition to exploit static and dynamic variations concurrently. Case studies on the Tennessee Eastman process and the boiler process are presented to illustrate the effectiveness of the proposed dynamic models over other existing dynamic methods.     

基于knee points的改进多目标人工蜂群算法

传统的人工蜂群算法（Artificial Bee Colony algorithm,ABC）及其在多目标上的扩展（Multi Objective Artificial Bee Colony algorithm,MOABC）存在着在高维、多峰函数情况下收敛速度变慢、后期容易陷入局部最优以及寻优精度丢失等问题。基于knee points提高收敛性和分布性的特点,设计了一种快速识别knee point的算法并将其应用到多目标人工蜂群算法中,提出了一种基于knee points的改进多目标人工蜂群算法（KnMOABC）。算法在迭代过程中考虑pareto支配关系的同时,优先选择knee point作为下一代个体,极大地增强了算法的收敛速度,同时,在knee point识别算法中加入自适应的策略以保持良好的分布性。实验结果表明,KnMOABC的性能优于三个最新的多目标人工蜂群对比算法。     

Exponential Stabilization of Switched Discrete‐Time Systems with All Unstable Modes

This paper studies the exponential stabilization of switched discrete‐time systems whose subsystems are unstable. A new sufficient condition for the exponential stability of the class of systems is proposed. The result obtained is based on the determination of a lower bound of the maximum dwell time by virtue of the multiple Lyapunov functions method. The key feature is that the given stability condition does not need the value of the Lyapunov function to uniformly decrease at every switching instant. An example is provided to illustrate the effectiveness of the proposed result.     

The binomial decomposition of OWA functions,the 2‐additive and 3‐additive cases in n dimensions

In the context of the binomial decomposition of ordered weighted averaging (OWA) functions, we investigate the constraints associated with the 2‐additive and 3‐additive cases in n dimensions. The 2‐additive case depends on one coefficient whose feasible region does not depend on the dimension n. On the other hand, the feasible region of the 3‐additive case depends on two coefficients and is explicitly dependent on the dimension n. This feasible region is a convex polygon with n vertices and n edges, which is strictly expanding in the dimension n. The orness of the OWA functions within the feasible region is linear in the two coefficients, and the vertices associated with maximum and minimum orness are identified. Finally, we discuss the 3‐additive binomial decomposition in the asymptotic infinite dimensional limit.     

Widlar微电流源的特性测量与建模研究

针对现有Widlar微电流源中电阻参数计算误差较大的问题,研究了一种基于Laguerre多项式拟合曲线的误差校正方法,并使用递推最小二乘法(RLS)对拟合曲线的模型参数进行递推计算,使得拟合曲线高精度拟合Widlar微电流源的非线性特性.研究结果表明,该方法具有很高的电阻参数计算精度,最大相对误差小于0.0963％,在Widlar微电流源设计领域具有重要的应用价值.     

Femtojoule-Power-Consuming Synaptic Memtransistor Based on Mott Transition of Multiphasic Vanadium Oxides

To realize the potential of Mott transition of multiphasic vanadium oxides (VO_x) for memory applications, the development of VO_x memtransistors on SiO₂ wafer is introduced. Through electrical characterizations, the volatile memory behaviors of the VO_x memtransistors are observed in both two- and three-terminal measurements. Their capacitive memory and resistive switching mechanisms are strongly related to the mixed VO_x/SiO₂ interface (called VSiO_x). VSiO_x supports the Mott transition in VO_x at low bias voltages (<0.5 V), leading to the low power consumption of the memtransistor. Moreover, the fast switching time (≈35 ns) and tunable memory retention with the synaptic functions (potentiation and depression) of the memtransistors (by using the gate and drain biases) are demonstrated. Overall, the findings open up major opportunities for constructing ultrafast and femto-joule power-consuming neuromorphic devices.     

Iterative computational approach to the solution of the Hamilton-Jacobi-Bellman-Isaacs equation in nonlinear optimal control

In this paper, iterative or successive approximation methods for the Hamilton-Jacobi-Bellman-Isaacs equations (HJBIEs) arising in both deterministic and stochastic optimal control for affine nonlinear systems are developed. Convergence of the methods are established under fairly mild assumptions, and examples are solved to demonstrate the effectiveness of the methods. However, the results presented in the paper are preliminary, and do not yet imply in anyway that the solutions computed will be stabilizing. More improvements and experimentation will be required before a satisfactory algorithm is developed.