随机Hopfield神经网络的定量分析

该文探讨了实际使用Hopfield神经网络(HNN)时噪声的影响。由于噪声的客观存在,我们首先证明了随机Hopfield神经网络(SHNN)轨道的期望关于时间是一致有界的。之后,为了实际设计神经网络的需要,我们对含有噪声的HNN和与其对应的一般HNN之间随机输入误差的估计进行了研究。利用所得的结论,我们可以对设计空间进行控制,使得所设计的网络满足我们希望获得的各种性能要求。     

Hopfield Neural Network Based on Estimation of Distribution for Two-Page Crossing Number Problem

This paper presents a Hopfield neural network (HNN) combined with estimation of distribution (EDA) for the two-page crossing number problem. In the proposed algorithm, once the network is trapped in local minima, the perturbation based on EDA can generate a new starting point for the HNN for further search, which is in a promising area characterized by a probability model and is not far away from the best solution found so far. The proposed algorithm can escape from local minima and further search better results. Simulation results show that the proposed algorithm is better than previous methods.      

随机petri网分析分组交换网中窗式流量控制机理

本文将随机Ｐｅｔｒｉ网与排队论相结合，对分组交换网中的窗式流量控制机理进行了描述与分析，建立了窗式流量控制机理的随机Ｐｅｔｒｉ网模型，并给出了随机Ｐｅｔｒｉ网模型的可达图及状态转移方程。为通信网的性能分析和评价提供了一种新的方法。     

A Hopfield neural-network-based dynamic channel allocation withhandoff channel reservation control

As channel allocation schemes become more complex and computationally demanding in cellular radio networks, alternative computational models that provide the means for faster processing time are becoming the topic of research interest. These computational models include knowledge-based algorithms, neural networks, and stochastic search techniques. This paper is concerned with the application of a Hopfield (1982) neural network (HNN) to dynamic channel allocation (DCA) and extends previous work that reports the performance of HNN in terms of new call blocking probability. We further model and examine the effect on performance of traffic mobility and the consequent intercell call handoff, which, under increasing load, can force call terminations with an adverse impact on the quality of service (QoS). To maintain the overall QoS, it is important that forced call terminations be kept to a minimum. For an HNN-based DCA, we have therefore modified the underlying model by formulating a new energy function to account for the overall channel allocation optimization, not only for new calls but also for handoff channel allocation resulting from traffic mobility. That is, both new call blocking and handoff call blocking probabilities are applied as a joint performance estimator. We refer to the enhanced model as HNN-DCA++. We have also considered a variation of the original technique based on a simple handoff priority scheme, here referred to as HNN-DCA+. The two neural DCA schemes together with the original model are evaluated under traffic mobility and their performance compared in terms of new-call blocking and handoff-call dropping probabilities. Results show that the HNN-DCA++ model performs favorably due to its embedded control for assisting handoff channel allocation     

Harmonic neural networks for on-line learning vector quantisation

A self-creating harmonic neural network (HNN) trained with a competitive algorithm effective for on-line learning vector quantisation is presented. It is shown that by employing dual resource counters to record the activity of each node during the training process, the equi-error and equi-probable criteria can be harmonised. Training in HNNs is smooth and incremental, and it not only achieves the biologically plausible on-line learning property, but it can also avoid the stability-plasticity dilemma, the dead-node problem, and the deficiency of the local minimum. Characterising HNNs reveals the great controllability of HNNs in favouring one criterion over the other, when faced with a must-choose situation between equi-error and equi-probable. Comparison studies on learning vector quantisation involving stationary and non-stationary, structured and non-structured inputs demonstrate that the HNN outperforms other competitive networks in terms of quantisation error, learning speed and codeword search efficiency     

Hopfield网络求解TSP的一种改进算法和理论证明

本文通过简化Ｈｏｐｆｉｌｅｄ神经网络求解问题的能量函数，提出了一种神经网络求解ＴＳＰ的改进算法，借助连接矩阵特征值的分析、从理论上证明了该算法保证获得ＴＳＰ有效解的原因。大量计算机模拟实验表明，该算法明显优于目前广泛应用的Ａｉｙｅｒ算法，具有收敛速度快、可避免无效解，易获得优化解等特点。     

A New Sequential Detection Based on Hopfield Neural Network in Frequency Selective Fading Channels

ＡＮｅｗＳｅｑｕｅｎｔｉａｌＤｅｔｅｃｔｉｏｎＢａｓｅｄｏｎＨｏｐｆｉｅｌｄＮｅｕｒａｌＮｅｔｗｏｒｋｉｎＦｒｅｑｕｅｎｃｙＳｅｌｅｃｔｉｖｅＦａｄｉｎｇＣｈａｎｎｅｌｓＷｅｎｇＪｉａｎｆｅｎｇ；ＢｉＧｕａｎｇｇｕｏ（ＳｏｕｔｈｅａｓｔＵｎｉｖｅｒｓ...     

Directional routing and scheduling for green vehicular delay tolerant networks

The vehicle delay tolerant networks (DTNs) make opportunistic communications by utilizing the mobility of vehicles, where the node makes delay-tolerant based “carry and forward” mechanism to deliver the packets. The routing schemes for vehicle networks are challenging for varied network environment. Most of the existing DTN routing including routing for vehicular DTNs mainly focus on metrics such as delay, hop count and bandwidth, etc. A new focus in green communications is with the goal of saving energy by optimizing network performance and ultimately protecting the natural climate. The energy–efficient communication schemes designed for vehicular networks are imminent because of the pollution, energy consumption and heat dissipation. In this paper, we present a directional routing and scheduling scheme (DRSS) for green vehicle DTNs by using Nash Q-learning approach that can optimize the energy efficiency with the considerations of congestion, buffer and delay. Our scheme solves the routing and scheduling problem as a learning process by geographic routing and flow control toward the optimal direction. To speed up the learning process, our scheme uses a hybrid method with forwarding and replication according to traffic pattern. The DRSS algorithm explores the possible strategies, and then exploits the knowledge obtained to adapt its strategy and achieve the desired overall objective when considering the stochastic non-cooperative game in on-line multi-commodity routing situations. The simulation results of a vehicular DTN with predetermined mobility model show DRSS achieves good energy efficiency with learning ability, which can guarantee the delivery ratio within the delay bound.     

On the energy-efficient organization and the lifetime of multi-hop sensor networks

Due to hardware, energy, cost and other physical constraints, sensor-based networks present various design, implementation and deployment challenges. An analytical model is presented to estimate and evaluate the node and network lifetime in a randomly deployed multi-hop sensor network. Based on this, we provide a procedure for the creation of an energy efficient sensor network organization, that attempts to extend the lifetime of the communication critical nodes, and as a result the overall network's operational lifetime.     

Improving local minima of columnar competitive model for TSPs

The columnar competitive model (CCM) has been recently proposed to solve the traveling salesman problem. This method performs much better than the original Hopfield network in terms of both the number and the quality of valid solutions. However, the local minima is still an open unsolved issue. This paper studies the performance of the CCM and aims to improve its local minima problem. The contributions of this paper are: 1) it proves by mathematics that the CCM is hardly to escape from local minimum states in general; 2) an alternate CCM is presented based on a modified energy function so as to enhance the capability of the original CCM; 3) a new algorithm is proposed by combining the alternative CCM with the original one, which enables the network to have lower energy level when trapped in local minima, this makes the network to reach the optimal or near-optimal state quickly; 4) Simulations are carried out to illustrate the performance of the proposed method.     

Abdominal organ segmentation using texture transforms and aHopfield neural network

Abdominal organ segmentation is highly desirable but difficult, due to large differences between patients and to overlapping grey-scale values of the various tissue types. The first step in automating this process is to cluster together the pixels within each organ or tissue type. We propose to form images based on second-order statistical texture transforms (Haralick transforms) of a CT or MRI scan. The original scan plus the suite of texture transforms are then input into a Hopfield neural network (HNN). The network is constructed to solve an optimization problem, where the best solution is the minima of a Lyapunov energy function. On a sample abdominal CT scan, this process successfully clustered 79-100% of the pixels of seven abdominal organs. It is envisioned that this is the first step to automate segmentation. Active contouring (e.g., SNAKE's) or a back-propagation neural network can then be used to assign names to the clusters and fill in the incorrectly clustered pixels.     

Reduced base model construction methods for stochastic activitynetworks

Reduced base model construction methods for stochastic activity networks are discussed. The basic definitions concerning stochastic networks are reviewed and the types of variables used in the construction process are defined. These variables can be used to estimate both transient and steady-state system characteristics. The construction operations used and theorems stating the validity of the method are presented. A procedure for generating the reduced base model stochastic process for a given stochastic activity network and performance variable is presented. Some examples which illustrate the method and demonstrate its effectiveness in reducing the size of a state space are presented     

Scaling Laws for Data-Centric Storage and Querying in Wireless Sensor Networks

We use a constrained optimization framework to derive scaling laws for data-centric storage and querying in wireless sensor networks. We consider both unstructured sensor networks, which use blind sequential search for querying, and structured sensor networks, which use efficient hash-based querying. We find that the scalability of a sensor network's performance depends upon whether the increase in energy and storage resources with more nodes is outweighed by the concomitant application-specific increase in event and query loads. We derive conditions that determine: 1) whether the energy requirement per node grows without bound with the network size for a fixed-duration deployment, 2) whether there exists a maximum network size that can be operated for a specified duration on a fixed energy budget, and 3) whether the network lifetime increases or decreases with the size of the network for a fixed energy budget. An interesting finding of this work is that three-dimensional (3D) uniform deployments are inherently more scalable than two-dimensional (2D) uniform deployments, which in turn are more scalable than one-dimensional (1D) uniform deployments.      

An architecture for mobile radio networks with dynamically changing topology using virtual subnets

An architecture adaptabie to dynamic topology changes in multi-hop mobile radio networks is described. The architecture partitions a mobile network into logically independent subnetworks. Network nodes are members of physical and virtual subnets and may change their affiliation with these subnets due to their mobility. Each node is allocated an address based on its current subnet affiliation. We observe-especially in large networks with random topology-that partitioning of the network may result in significantly more balanced load than in one large multi-hop network, an attribute that can significantly improve the network's performance. The architecture is highly fault-tolerant, has a relatively simple location updating and tracking scheme, and by virtue of its load balancing feature, typically achieves a network with relatively high throughput and low delay. The addressing method, logical topology, mobility management and routing procedure are described, and network performance is evaluated.     

Spectroscopy and hybrid neural network analysis

This paper reviews the current use of spectroscopy and related instrumentation in chemical analysis. Advancements in digital signal processing technology are making it possible to improve the sensitivity and accuracy of analytical instruments without expensive upgrading of instrument hardware. A hybrid neural network (HNN) is described that can perform nonlinear signal analysis. The HNN approach combines the simple data reduction capability of conventional linear signal processing algorithms with the adaptive learning and recognition ability of a multilayer nonlinear neural network architecture. A number of examples show the rise of the HNN for environmental monitoring and real-time process control     

An efficient bandwidth allocation scheme for hierarchical cellular networks with energy harvesting: an actor-critic approach

In this paper, we propose an efficient bandwidth allocation strategy for multiclass services in hierarchical cellular networks that consist of an operation controller, several small-cell base stations (SBSs), and a number of mobile users. Each SBS is equipped with a finite-capacity battery that is regularly recharged by a solar harvester. We aims to find the optimal bandwidth allocation policy in order to enhance the network performance in terms of user satisfaction and energy efficiency under energy harvesting and bandwidth sharing constraints. Since the arrivals of harvested energy and traffic requests are unknown due to users’ mobility and stochastic request generation, it is necessary to design a learning framework for the controller in order to predict these dynamics through interaction with the environment. For this purpose, we first formulate the resource allocation problem as the framework of a Markov decision process, and then, we employ an actor-critic algorithm to find the optimal policy under which the controller can effectively allocate the limited bandwidth to the SBSs for their data transmissions. We evaluate the performance of the proposed scheme through comprehensive simulations with different settings, and show that the proposed bandwidth allocation scheme can enhance the network’s performance in the long run.     

数字电路的最优神经网络模型及建立方法

本文研究电路的最优神经网络模型,获得了对任意结构的多输入多输出逻辑电路,都存在一种最优神经网络能表征电路的逻辑功能,通过求解一个线性方程组可以得到这种神经网络的结构.文中也给出了多输入基本门电路的最优神经网络结构及其能量函数的一般表达式.     

Energy-aware wireless networking with directional antennas: the case of session-based broadcasting and multicasting

We consider ad hoc wireless networks that use directional antennas and have limited energy resources. To explore quantitatively the advantage offered by the use of directional antennas over the case of omnidirectional antennas, we consider the case of connection-oriented multicast traffic. Building upon our prior work on multicasting algorithms, we introduce two protocols that exploit the use of directional antennas and evaluate their performance, We observe significant improvement with respect to the omnidirectional case, in terms of both energy efficiency and network lifetime. Additionally, we show that further substantial increase in the network's lifetime can be achieved by incorporating a simple measure of a node's residual energy into the node's cost function.     

Least-Squares Design of FIR Filters Based on a Compacted Feedback Neural Network

The design of finite-impulse response (FIR) filters can be performed by using neural networks by formulating the objective function to a Lyapunov energy function. Focusing on this goal, the authors present an improved structure of a feedback neural network to implement the least-squares design of FIR filters. In addition to using the closed-form expressions for the synaptic weight matrix and the bias parameter of the Hopfield neural network (HNN), the proposed approach can achieve a notable reduction both in the amount of computation required and hardware complexity compared to the previous neural-based method. Simulation results indicate the effectiveness of the proposed approach