G-networks with resets

Gelenbe networks (G-networks) are product form queuing networks which, in addition to ordinary customers, contain unusual entities such as “negative customers” which eliminate normal customers and “triggers” which move other customers from some queue to another. These models have generated much interest in the literature since the early 1990s. The present paper discusses a novel model for a reliable system composed of N unreliable systems, which can hinder or enhance each other’s reliability. Each of the N systems also tests other systems at random; it is able to reset another system if it is itself in working condition and discovers that the other system has failed, so that the global reliability of the system is enhanced. This paper shows how an extension of G-networks that includes novel “reset” customers can be used to model this behavior. We then show that a general G-network model with resets has product form, and prove existence and uniqueness of its solution.     

G-Networks: Development of the Theory of Multiplicative Networks

This is a review on G-networks, which are the generalization of the Jackson and BCMP networks, for which the multi-dimensional stationary distribution of the network state probabilities is also represented in product form. The G-networks primarily differ from the Jackson and BCMP networks in that they additionally contain a flow of the so-called negative customers and/or triggers. Negative customers and triggers are not served. When a negative customer arrives at a network node, one or a batch of positive (ordinary) customers is killed (annihilated, displaced), whereas a trigger displaces a positive customer from the node to some other node. For applied mathematicians, G-networks are of great interest for extending the multiplicative theory of queueing networks and for practical specialists in modeling computing systems and networks and biophysical neural networks for solving pattern recognition and other problems.     

Decomposition of Queueing Networks with Dependent Service and Negative Customers

Queueing networks with negative customers (G-networks) and dependent service at different nodes are studied. Every customer arriving at the network is defined by a set of random parameters: his route over the network (a sequence of nodes visited by the customers), route length, and volume and service length of the customer at every stage of the route. For G-networks, which are the analogs of BCMP-networks, the multidimensional stationary distribution of the network state probabilities is shown to be representable in product form.     

G-networks with synchronised arrivals

We consider G-networks in which customers arrive simultaneously in several queues. We denote this new signal as synchronised arrivals. Under some conditions on the arrivals on the boundary of the state space, and the ergodicity condition, we prove that these networks have a product-form steady-state distribution. We show the link between this new signal and the positive signals introduced by Chao, Miyazawa and Pinedo.     

Introduction to the Special Issue on G-Networks and the Random Neural Network

This special issue includes seven papers on G-networks and the Random Neural Network (RNN), approximately twenty years after these models were first introduced. We summarise the differences between the research trends related to these two branches of the same initial model, and point to applications that are less known to the Performance community and suggest significant questions related to G-networks and the RNN that are fruitful research directions.     

On the solution of a rational matrix equation arising in G-networks

We consider the problem of solving a rational matrix equation arising in the solution of G-networks. We propose and analyze two numerical methods: a fixed point iteration and the Newton–Raphson method. The fixed point iteration is shown to be globally convergent with linear convergence rate, while the Newton method is shown to have a local convergence, with quadratic convergence rate. Numerical experiments show the effectiveness of the proposed methods.     

Network routing control with G-networks

The aim of this paper is to detail a control scheme for packet computer networks whose purpose is to minimise a quality-of-service oriented performance metric by re-routing the traffic. The model is based on G-networks with triggered customer movement to represent traffic re-routing, and on a gradient descent based optimisation algorithm. The model and the algorithm are presented and we show that the gradient descent algorithm is of computational complexity O(N³) where N is the number of nodes in the packet network. Via the use of multiple classes of normal traffic and multiple classes of triggers, our approach allows one not only to evaluate the effect of the control, but also to incorporate the overhead that the control traffic will induce, and the consequences of the delays or possible losses of the control traffic. Similarly, these effects will naturally be incorporated when one considers both the impact of the control traffic on the cost function, and the details of this control traffic in the control algorithm itself.     

Separable equilibrium state probabilities via time reversal in Markovian process algebra

The reversed compound agent theorem (RCAT) is a compositional result that uses Markovian process algebra (MPA) to derive the reversed process of certain interactions between two continuous time Markov chains at equilibrium. From this reversed process, together with the given, forward process, the joint state probabilities can be expressed as a product-form, although no general algorithm has previously been given. This paper first generalises RCAT to multiple (more than two) cooperating agents, which removes the need for multiple applications and inductive proofs in cooperations of an arbitrary number of processes. A new result shows a simple stochastic equivalence between cooperating, synchronised processes and corresponding parallel, asynchronous processes. This greatly simplifies the proof of the new, multi-agent theorem, which includes a statement of the desired product-form solution itself as a product of given state probabilities in the parallel components. The reversed process and product-form thus derived rely on a solution to certain rate equations and it is shown, for the first time, that a unique solution exists under mild conditions—certainly for queueing networks and G-networks.     

神经网络集成在图书剔旧分类中的应用

在分析图书剔旧工作的基础上，指出用智能的方法解决图书剔旧问题的必要性。提出了可以用神经网络集成技术来解决该问题，并给出一种动态构建神经网络集成的方法，该方法在训练神经网络集成成员网络时不仅调整网络的连接权值，而且动态地构建神经网络集成中各成员神经网络的结构，从而在提高单个网络精度的同时，增加了各网络成员之间的差异度，减小了集成的泛化误差。实验证明该方法可以有效地用于图书剔旧分类。     

基于细胞神经网络的灰度图像负片算法

提出了一种基于细胞神经网络的灰度图像负片算法。根据细胞神经网络高速并行的特点。提出并设计了单层细胞神经网络负片模板用于灰度图像的负片处理。为细胞神经网络在图像处理领域中的应用提供了一种优良的算法。实验证明了，该算法对灰度图像负片处理的有效性。     

基于人工神经网络的负相关学习研究

人工神经网络集成技术是神经计算技术的一个研究热点,在许多领域中已经有了成熟的应用.神经网络集成是用有限个神经网络对同一个问题进行学习,集成在某输入示例下的输出由构成集成的各神经网络在该示例下的输出共同决定.负相关学习法是一种神经网络集成的训练方法,它鼓励集成中的不同个体网络学习训练集的不同部分,以使整个集成能更好地学习整个训练数据.改进的负相关学习法是在误差函数中使用一个带冲量的BP算法,给合了原始负相关学习法和带冲量的BP算法的优点,使改进的算法成为泛化能力强、学习速度快的批量学习算法.     

基于负相关神经网络集成算法及其应用的研究

传统的神经网络集成中各子网络之间的相关性较大,从而影响集成的泛化能力.为此,提出用负相关学习算法来训练神经网络集成,以增加子网络间的差异度,从而提高集成的泛化能力.并将基于负相关学习法的神经网络集成应用于中医舌诊诊断,以肝病病证诊断进行仿真.实验结果表明:基于负相关学习法的神经网络集成比单个子网和传统神经网络集成更能有效地提高其泛化能力.因此,基于负相关神经网络集成算法的研究是可行的、有效的.     

基于阴性选择算法的电力负荷异常检测与调整

提出用阴性选择算法对电力负荷数据进行异常检测及调整,其中自我定义为正常负荷数据模式,而非我则为偏差超过一定阈值的负荷数据模式.先用Kohonen网对日负荷曲线进行聚类,产生各类的特征曲线,并将其编码作为自我集S;再用随机方法产生检测器集R.被检负荷与R集进行匹配审查,以判断正常与否,并对检测出的异常负荷用自我集S进行调整.实验结果表明,用阴性选择算法对负荷数据进行异常检测和调整取得令人满意的结果.     

Negative correlation in incremental learning

Negative Correlation Learning (NCL) has been successfully applied to construct neural network ensembles. It encourages the neural networks that compose the ensemble to be different from each other and, at the same time, accurate. The difference among the neural networks that compose an ensemble is a desirable feature to perform incremental learning, for some of the neural networks can be able to adapt faster and better to new data than the others. So, NCL is a potentially powerful approach to incremental learning. With this in mind, this paper presents an analysis of NCL, aiming at determining its weak and strong points to incremental learning. The analysis shows that it is possible to use NCL to overcome catastrophic forgetting, an important problem related to incremental learning. However, when catastrophic forgetting is very low, no advantage of using more than one neural network of the ensemble to learn new data is taken and the test error is high. When all the neural networks are used to learn new data, some of them can indeed adapt better than the others, but a higher catastrophic forgetting is obtained. In this way, it is important to find a trade-off between overcoming catastrophic forgetting and using an entire ensemble to learn new data. The NCL results are comparable with other approaches which were specifically designed to incremental learning. Thus, the study presented in this work reveals encouraging results with negative correlation in incremental learning, showing that NCL is a promising approach to incremental learning.

Clustering in diffusively coupled networks

This paper shows how different mechanisms may lead to clustering behavior in connected networks consisting of diffusively coupled agents. In contrast to the widely studied synchronization processes, in which the states of all the coupled agents converge to the same value asymptotically, in the cluster synchronization problem studied in this paper, we require all the interconnected agents to evolve into several clusters and each agent only to synchronize within its cluster. The first mechanism is that agents have different self-dynamics, and those agents having the same self-dynamics may evolve into the same cluster. When the agents’ self-dynamics are identical, we present two other mechanisms under which cluster synchronization might be achieved. One is the presence of delays and the other is the existence of both positive and negative couplings between the agents. Some sufficient and/or necessary conditions are constructed to guarantee n-cluster synchronization. Simulation results are presented to illustrate the effectiveness of the theoretical analysis.     

A neural networks-based negative selection algorithm in fault diagnosis

Inspired by the self/nonself discrimination theory of the natural immune system, the negative selection algorithm (NSA) is an emerging computational intelligence method. Generally, detectors in the original NSA are first generated in a random manner. However, those detectors matching the self samples are eliminated thereafter. The remaining detectors can therefore be employed to detect any anomaly. Unfortunately, conventional NSA detectors are not adaptive for dealing with time-varying circumstances. In the present paper, a novel neural networks-based NSA is proposed. The principle and structure of this NSA are discussed, and its training algorithm is derived. Taking advantage of efficient neural networks training, it has the distinguishing capability of adaptation, which is well suited for handling dynamical problems. A fault diagnosis scheme using the new NSA is also introduced. Two illustrative simulation examples of anomaly detection in chaotic time series and inner raceway fault diagnosis of motor bearings demonstrate the efficiency of the proposed neural networks-based NSA.     

Automatic extraction of eye and mouth fields from a face image using eigenfeatures and multilayer perceptrons

This paper presents a novel algorithm for the extraction of the eye and mouth (facial features) fields from 2-D gray-level face images. The fundamental philosophy is that eigenfeatures, derived from the eigenvalues and eigenvectors of the binary edge data set constructed from the eye and mouth fields, are very good features to locate these fields efficiently. The eigenfeatures extracted from the positive and negative training samples of the facial features are used to train a multilayer perceptron whose output indicates the degree to which a particular image window contains an eye or a mouth. It turns out that only a small number of frontal faces are sufficient to train the networks. Furthermore, they lend themselves to good generalization to non-frontal pose and even other people's faces. It has been experimentally verified that the proposed algorithm is robust against facial size and slight variations of pose.     

Delay-dependent stability analysis for impulsive neural networks with time varying delays

In this paper, the global exponential stability and global asymptotic stability of the neural networks with impulsive effect and time varying delays is investigated. By using Lyapunov–Krasovskii-type functional, the quality of negative definite matrix and Cauchy criterion, we obtain the sufficient conditions for global exponential stability and global asymptotic stability of such model, in terms of linear matrix inequality (LMI), which depend on the delays. Two examples are given to illustrate the effectiveness of our theoretical results.     

Combining features of negative correlation learning with mixture of experts in proposed ensemble methods

Both theoretical and experimental studies have shown that combining accurate neural networks (NNs) in the ensemble with negative error correlation greatly improves their generalization abilities. Negative correlation learning (NCL) and mixture of experts (ME), two popular combining methods, each employ different special error functions for the simultaneous training of NNs to produce negatively correlated NNs. In this paper, we review the properties of the NCL and ME methods, discussing their advantages and disadvantages. Characterization of both methods showed that they have different but complementary features, so if a hybrid system can be designed to include features of both NCL and ME, it may be better than each of its basis approaches. In this study, two approaches are proposed to combine the features of both methods in order to solve the weaknesses of one method with the strength of the other method, i.e., gated-NCL (G-NCL) and mixture of negatively correlated experts (MNCE). In the first approach, G-NCL, a dynamic combiner of ME is used to combine the outputs of base experts in the NCL method. The suggested combiner method provides an efficient tool to evaluate and combine the NCL experts by the weights estimated dynamically from the inputs based on the different competences of each expert regarding different parts of the problem. In the second approach, MNCE, the capability of a control parameter for NCL is incorporated in the error function of ME, which enables the training algorithm of ME to efficiently adjust the measure of negative correlation between the experts. This control parameter can be regarded as a regularization term added to the error function of ME to establish better balance in bias–variance–covariance trade-offs and thus improves the generalization ability. The two proposed hybrid ensemble methods, G-NCL and MNCE, are compared with their constituent methods, ME and NCL, in solving several benchmark problems. The experimental results show that our proposed methods preserve the advantages and alleviate the disadvantages of their basis approaches, offering significantly improved performance over the original methods.     

A novel frequency-selective metamaterial to improve helix antenna

A novel frequency-selective metamaterial with negative permittivity and permeability for improving directivity and gain of a helix antenna is presented in this paper.The proposed metamaterial is composed of two Z-shape resonators printed on opposite sides of a dielectric substrate.Two forms of multilayered cells are found to be suitable for antennas and waveguides applications.In addition,a new method of designing a metamaterial-based helix antenna is presented with high directivity and gain.A comparison on radiation properties is given between the conventional and the new metamaterial-based helix antennas.Two comparisons on radiation properties are performed:(1) the effect of proposed Z-structure on monopole,dipole,and helix antennas;(2) the effect of OE3,split-ring resonator (SRR),and proposed Z-structure unit cells on the performance of helix antennas.The results show improvement of parameters such as directivity,gain,and radiation power of the new metamaterial-based helix antenna.Therefore,the combination of Z-structure with the helix antenna shows the best performance.