Generalized irreducible memory function

In a previous paper we presented a general scheme to reduce the memory function in nonequilibrium statistical physics for purely dissipative cases with detailed balance. Here we simplify and generalize further this scheme to include cases other than purely dissipative ones. As an illustration we discuss simple dense fluids.     

A memory device sustained by noise

We propose a memory device consisting of a ring of two bistable coupled oscillators. We show that this system is capable of storing a single bit and its performance improves with noise, in agreement with previous experimental results obtained with discrete bistable electronic circuits. The stored bit can be retrieved asynchronously and, after a certain synchronization time, the probability of error does not depend on the interrogated oscillator. Memory persistence is shown to be maximized for the same noise range that both minimizes the probability of error and ensures synchronization.     

Dynamics of moving agents with memory

The problem of emergence of collective behavior in a system of two dimensional interacting moving agents is considered. The individual agent architecture is hybrid in the sense that the response is reactive but the agent is capable to choose between two different rules in a probabilistic way: one rule depends on the information from the environment and the other depends on the information imprinted in the agents memory. The dynamics of the system is studied through the characterization of the equilibrium state and the non-equilibrium properties of the system.     

Magnetic memory effect in chelated zero valent iron nanoparticles

We report the study of nonequilibrium magnetic behavior of air stable zero valent iron nanoparticles synthesized in presence of N-cetyl-N,N,N-trimethyl ammonium bromide chelating agent. X-ray photoelectron spectroscopy study has suggested the presence of iron oxides on nZVI surfaces. Zero-field-cooled and field-cooled magnetization measurements have been carried out at 20–300 K and 100 Oe. For field-cooled measurements with 1 h stops at 200, 100 and 50 K when compared with the warming cycle, we found the signature of magnetic memory effect. A study of magnetic relaxation at the same temperatures shows the existence of two relaxation times.     

Human memory retrieval as Lévy foraging

When people attempt to recall as many words as possible from a specific category (e.g., animal names) their retrievals occur sporadically over an extended temporal period. Retrievals decline as recall progresses, but short retrieval bursts can occur even after tens of minutes of performing the task. To date, efforts to gain insight into the nature of retrieval from this fundamental phenomenon of semantic memory have focused primarily upon the exponential growth rate of cumulative recall. Here we focus upon the time intervals between retrievals. We expected and found that, for each participant in our experiment, these intervals conformed to a Lévy distribution suggesting that the Lévy flight dynamics that characterize foraging behavior may also characterize retrieval from semantic memory. The closer the exponent on the inverse square power-law distribution of retrieval intervals approximated the optimal foraging value of 2, the more efficient was the retrieval. At an abstract dynamical level, foraging for particular foods in one's niche and searching for particular words in one's memory must be similar processes if particular foods and particular words are randomly and sparsely located in their respective spaces at sites that are not known a priori. We discuss whether Lévy dynamics imply that memory processes, like foraging, are optimized in an ecological way.     

Disappearance of time integrals of exact memory functions in time-convolution generalized master equations

Memory functions in time-convolution Generalized Master Equations (GME) for probabilities of finding a general system (interacting by a general coupling with a true thermodynamic bath) in individual states are considered without resorting to any approximation. After taking the thermodynamic bath limit, time integrals from zero to infinite times of the memories are considered. It is argued that these integrals entering, e.g., the usual naive Markov approximation converting GME the Pauli master (PME) equations are exactly zero. This implies long-time tails of memories (unobtainable by perturbational expansions) and slower-than-exponential long-time asymptotics of relaxation.     

Study of nuclear tracks on different polycarbonate foils and their feasibility for use in personnel monitoring

CR-39 (PADC) a Solid State Nuclear Track Detector., is used as a routine personnel monitor. 1800 workers are being monitored quarterly for fast neutron using CR-39 foils. These foils procured from Pershore Mouldings, UK are very expensive and indigenous development will make the foils cost effective. The aim of this paper is to find a suitable alternative to the imported CR-39 foils for use in personnel monitoring. The foils from three different manufacturers have been compared with the CR-39 foils from Pershore Moulding, UK, presently in use. Out of the three, only sample no 1 is promising. It has a background and sensitivity comparable with CR-39 presently used. The sample 2 is CR-39 being developed in India, has a relatively high background and poor sensitivity. Efforts are being made to improve the quality of this sample. The sample 3 was a poly carbonate from local manufacturer which produced very few tracks and the standard deviation of track counts was very large and hence not useful for Personnel Monitoring.      

Kernel spectral clustering with memory effect

Evolving graphs describe many natural phenomena changing over time, such as social relationships, trade markets, metabolic networks etc. In this framework, performing community detection and analyzing the cluster evolution represents a critical task. Here we propose a new model for this purpose, where the smoothness of the clustering results over time can be considered as a valid prior knowledge. It is based on a constrained optimization formulation typical of Least Squares Support Vector Machines (LS-SVM), where the objective function is designed to explicitly incorporate temporal smoothness. The latter allows the model to cluster the current data well and to be consistent with the recent history. We also propose new model selection criteria in order to carefully choose the hyper-parameters of our model, which is a crucial issue to achieve good performances. We successfully test the model on four toy problems and on a real world network. We also compare our model with Evolutionary Spectral Clustering, which is a state-of-the-art algorithm for community detection of evolving networks, illustrating that the kernel spectral clustering with memory effect can achieve better or equal performances.     

More memory under evolutionary learning may lead to chaos

We show that an increase of memory of past strategy performance in a simple agent-based innovation model, with agents switching between costly innovation and cheap imitation, can be quantitatively stabilising while at the same time qualitatively destabilising. As memory in the fitness measure increases, the amplitude of price fluctuations decreases, but at the same time a bifurcation route to chaos may arise. The core mechanism leading to the chaotic behaviour in this model with strategy switching is that the map obtained for the system with memory is a convex combination of an increasing linear function and a decreasing non-linear function.     

Multi-layer graphene membrane based memory cell

The scheme and operational principles of the nanoelectromechanical memory cell based on the bending of a multi-layer graphene membrane by the electrostatic force are proposed. An analysis of the memory cell total energy as a function of the memory cell sizes is used to determine the sizes corresponding to a bistable memory cell with the conducting ON and non-conducting OFF states and to calculate the switching voltage between the OFF and ON states. It is shown that a potential barrier between the OFF and ON states is huge for practically all sizes of a bistable memory cell which excludes spontaneous switching and allows the proposed memory cell to be used for long-term archival storage.     

Generalized transport model for phase transition with memory

A general model for phenomenological transport in phase transition is derived, which extends Jäckle and Frisch model of phase transition with memory and the Cahn–Hilliard model. In addition to including interfacial energy to account for the presence of interfaces, we introduce viscosity and relaxation contributions, which result from incorporating memory effect into the driving potential. Our simulation results show that even without interfacial energy term, the viscous term can lead to transient diffuse interfaces. From the phase transition induced hysteresis, we discover different energy dissipation mechanism for the interfacial energy and the viscosity effect. In addition, by combining viscosity and interfacial energy, we find that if the former dominates, then the concentration difference across the phase boundary is reduced; conversely, if the interfacial energy is greater then this difference is enlarged.     

Experimental investigation of noise-assisted information transmission and storage via stochastic resonance

We present experimental results on the information transmission and storage via stochastic resonance in circuits designed and built around Schmitt triggers (STs). First, we investigate the performance of a transmission line comprised of five STs and show it to exhibit stochastic resonance. Each ST in the line is fed with white Gaussian noise, and the first ST is driven by a non-return-to-zero pseudo-random bit sequence with sub-threshold amplitude. Parameters such as bit error rate (Q-factor) are measured (calculated) and shown to exhibit a minimum (maximum) for an optimum amount of noise. Interestingly, we find that system performance degrades with the number of STs as if the system were linear and impaired only by additive Gaussian noise. We then propose and build a 1-bit storage device based on two STs in a loop configuration. We demonstrate that such a system is capable of storing one bit of information only in the presence of noise, and that there is a regime where the efficiency of such a device increases with increasing noise.Our results point to the feasibility of building ‘blocks’ that can transmit, store and eventually process information, whose performance is not only robust against noise, but can actually benefit from it.     

多核环境下内存数据库并发调度技术优化研究

对多核环境下内存数据进行并发调度,可以减少计算机宕机次数和数据切换时时间,提高数据并发调度精度,增加数据操作平稳性。当前的内存数据并发调度方法是利用PrebuiltTrigger对内存数据进行并发调度,在调度过程中,没有设定具体的内存数据调度目标,导致内存数据库中的数据因此错乱无序,存在数据并发调度精度低的问题。为此,提出一种基于Linux的多核环境下内存数据并发调度优化方法。该方法首先采用IACT算法对影响调度进行的数据和内存数据库中相似或重复数据进行清洗,然后以清洗的数据为基础,利用启发式算法对其进行数据特征选取,依据多属性决策理论对内存数据并发调度的最优路径属性权重集合进行计算,以其结果为依据,计算调度最优路径的偏差值,最后利用最小偏差值,建立调度最优路径线性规划模型,对每条调度路径的综合决策属性值进行排序,由此得到调度的最优路径,完成对多核环境下内存数据的并发调度。实验结果证明,所提方法可以对多核环境下内存数据进行高效率地并发调度,提高了数据调度精度,增加了内存数据的可循环利用性,为低开销的内存数据调度提供了支撑。     

On the memory functions in simple classical liquids

The relation between the two memory function formalisms for correlation functions in classical liquids is discussed. It is found that the kinetic equation formalism of Duderstadt and Akcasu with a simple exponential memory function can account for the double Gaussian form of the memory function in the generalized-hydrodynamics approach. The former therefore gives reasonably good results for the coherent scattering function S(k, ), as is shown for the case of liquid Rb at 315 K in the range 1.25 k 5.5 Å^–.     

基于RS485总线的存储器网络化技术研究

传统网络化测试技术存在网络化测试精准度低、不能对全部数据进行测试、数据属性测试不明确等问题,为此,基于RS485串行连接框图和存储器结构对网络化测试技术展开研究。采用信息融合技术中的信息融合算法得出测试估计误差方阵,完善传统技术中精准度低的问题。利用定义模型和动态模型有限内存的处理技术,对整个存储内存进行访问,从而对全部数据进行测试。采用分布式计算技术,利用二元制算法测试存储器中数据的质量,从而确定数据的属性。通过实验得出结论,本文测试技术不但能够确定数据属性,还能提高网络化测试精准度,并对全部数据进行测试,为我国未来网络化测试技术奠定了坚实的基础。     

Realizing reliable logic and memory function with noise-assisted Schmitt trigger circuits

Schmitt trigger is a comparator circuit with hysteresis, implemented by applying positive feedback to a noninverting input. Recently, it has been verified that noise under certain condition can help nonlinear circuit to realize reliable logical stochastic resonance (LSR). In this letter, we demonstrate that, without any modifications, Schmitt trigger can operate flexibly both as logic operation and latch operation in an optimal noise range. In traditional Schmitt trigger, the hysteresis is designed for resisting noise. In the new design, the hysteresis tends to enhance the LSR effect in Schmitt trigger with the help of noise and helps the system to perform better at higher noise levels. Thus, our results extend the application scope and generality of the traditional Schmitt trigger.     

Nonvolatile graphene nanoflake shuttle memory

We present the possible schematics of nanoscale graphene-flake shuttle-memory, discuss its basic principles, and investigate the energetic and the dynamic properties via classical molecular dynamics simulations. Graphite flake surface, where the nanoscale graphene-flake’s stage of activity takes place, has a dumbbell shape, and a nanoscale graphene-flake is placed on its surface. The van der Waals interactions between the graphene-flake and the patterned graphite make the bistable potential energy wells in the larger surface area regions of the patterned graphite, and then the graphene-flake keeps its seat on one of the bistable positions, which is the place where the binary data is archived. Since the movable graphene-flake can be also shuttled between the bistable states through the narrow passageway, this proposed nano-graphene-flake shuttle-memory can be utilized from nanoscale switchable nonvolatile memory.     

基于uClinux的嵌入式水声信号采集存储系统设计

为了减少水声学研究和水声工程设计试验的时间、成本等,针对水声信号和系统工作环境的特点,设计并实现了一种基于嵌入式uClinux操作系统的水声信号采集存储系统。发挥了处理器对外部设备优秀的控制能力和uClinux操作系统出色的稳定性两大优势,使系统稳定、可靠、低功耗、便捷地工作。通过海试验证了数据的有效性,证明系统具有实用价值。     

Investigation of liquid lithium relaxation time by means of the memory function formalism

The temporal characteristics of relaxation processes in liquid lithium extracted from the results of inelastic neutron scattering experiments, are analyzed with the use of the memory function formalism.     

Novel memory efficient LDPC decoders for beyond 5G

Beyond-5G wireless networks are expected to gain a excellent trade-off among computational accuracy, latency, and efficient use of available resources. This poses a significant challenge to the channel decoder. In this paper, a novel memory efficient algorithm for decoding Low-Density Parity-Check (LDPC) codes is proposed with a view to reduce the implementation complexity and hardware resources. The algorithm, called Check Node Self-Update (CNSU) algorithm, is based on layered normalized min-sum (LNMS) decoding algorithm while utilizing iteration parallel techniques to integrate both Variable Nodes (VNs) message and A-Posterior Probability(APP) message into the Check Nodes (CNs) message, which eliminates memories of both the VNs message and the APP message as well as updating module of APP message in CNs unit. Based on the proposed CNSU algorithm, design of partially parallel decoder architecture and serial simulations followed by implementation on the Stratix II EP2S180 FPGA are presented. The results show that the proposed algorithm and implementation bring a significant gain in efficient using of available resources, include reducing hardware memory resources and chip area while keeping the benefit of bit-error-rate (BER) performance and speeding up of convergence with LNMS, which are beneficial to apply in Beyond-5G wireless networks.