The use of a synchronizer yields the maximum computation rate in distributed networks

We study the performance of networks whose operation is controlled by a simple synchronizer. In a previous paper we analyzed the performance of networks with negligible transmission delay. It was shown that full speed is achieved, for any wake-up pattern, by letting the network run free, without the use of a “firing-squad” mechanism or a scheduler. In this paper we investigate the effect of fixed delays in the communication channels on the performance of a network in which there is a global clock, but there is no global start-up signal. We show that here, too, the maximum rate of computation is always reached, just by using the synchronizer and letting the network run free. To a certain extent, the wake-up pattern may influence the length of the transitory stage and the periodicity of the steady state, but not the ultimate rate. In any case, the length of the transitory stage is bounded polynomially, and the bound is tight, while the period may be of exponential length. An extended abstract of this paper appeared in STOC ’90. Part of this work was done while the authors visited the Computer Science Program, University of Texas at Dallas, Richardson, TX, USA. The first author is now visiting Bell Laboratories, Lucent Technologies, Murray Hill, NJ, USA, and was supported by the Fund for the Promotion of Research at the Technion. The second author’s current address is Instituto de Matematicas — U.N.A.M., Ciudad Universitaria, D.F. 04510, Mexico. rajsbaum@servidor.unam.mx     

Writing, technologies, and the fifth canon

Andrea Lunsford's keynote address to the 2005 Computers and Writing Conference at Stanford University, Palo Alto, California, expands the definition of writing to include epistemic, multivocal, multimodal, and multimediated practices in the computers and writing classroom. The article describes the development and piloting of a new undergraduate course in Stanford's Program in Writing and Rhetoric that applies these concepts to the undergraduate composition process. The address closes with a challenge to create classroom experiences that allow students to compose in “the most compelling discursive modalities of their generation”.     

Artificial neural networks controlled fast valving in a powergeneration plant

This paper presents an artificial neural-network-based controller to realize the fast valving in a power generation plant. The backpropagation algorithm is used to train the feedforward neural networks controller. The hardware implementation and the test results of the controller on a physical pilot-scale power plant setup are described in detail. Compared with the conventional fast valving methods applied to the same system, test results both with the computer simulation and on a physical pilot-scale power plant setup demonstrate that the artificial neural network controller has satisfactory generalization capability, reliability, and accuracy to be feasible for this critical control operation.     

异步时钟域的亚稳态问题和同步器

相较纯粹的单一时钟的同步电路设计,设计人员更多遇到的是多时钟域的异步电路设计.因此,异步电路设计在数字电路设计中的重要性不言而喻.本文主要就异步设计中涉及到的亚稳态问题,作简要介绍,并提出常用的解决办法--即同步器的使用.     

A study on shuffle, stopwatches and independently evolving clocks

We show that stopwatch automata are equivalent with timed shuffle expressions, an extension of timed regular expressions with the shuffle operation. Since the emptiness problem is undecidable for stopwatch automata, and hence also for timed shuffle expressions, we introduce a decidable subclass of stopwatch automata called partitioned stopwatch automata. We give for this class an equivalent subclass of timed shuffle expressions and investigate closure properties by showing that partitioned stopwatch automata are closed under union, concatenation, star, shuffle and renaming, but not under intersection. We also show that partitioned stopwatch automata are equivalent with distributed time-asynchronous automata, which are asynchronous compositions of timed automata in which time may evolve independently.     

一种基于跨层功率控制的Ad hoc网络路由算法*

在无线Ad Hoc网络路由协议中引入功率控制不但可以降低网络能量消耗,同时还能改善网络的吞吐量、投递率等性能,已成为当前Ad Hoc网络的一个研究热点.本文提出了一种基于跨层功率控制的按需路由算法CPC-AODV(Cross-layer Power Control Ad hoc On-demand Distance Vector).算法按需建立多个不同功率级的路由,节点选择到目的节点最小功率级的路由来传递分组,并对网络层的数据分组、路由分组和MAC层控制帧的传输采用不同功率控制策略来降低能量消耗.仿真结果表明:算法有利于降低通信能量开销,延长网络寿命,提高网络投递率及改善网络时延.     

A self-stabilizing algorithm to synchronize digital clocks in a distributed system

We have proposed a self-stabilizing algorithm to synchronize multiple digital clocks in a distributed system; whenever any of the clock values gets out of synchronization for any reason, the algorithm is automatically invoked and the system is brought back to a legitimate state in finite time.     

一类受控闭排队网络基于性能势的最优性方程

研究一类受控闭排队网络系统的性能优化问题. 文章引进了两个基本概念: 折扣代价α 性能势和平均代价性能势, 并且讨论了这两个性能势之间的一个关系式. 在一般的假设条件下, 我们应用性能势的基本性质直接建立了无限时间水平平均代价模型的最优性方程, 并且证明了在紧致集上最优解的存在性. 最后给出了一个策略优化的迭代算法并通过一个实际算例以说明该算法的效果.     

Maximizing connected coverage via controlled actor relocation in wireless sensor and actor networks

Wireless sensor and actor networks (WSANs) have recently emerged with the idea of combining wireless sensor networks (WSNs) and mobile ad hoc networks (MANETs). In addition to resource constrained sensors, resource rich and mobile actor nodes are employed in WSANs. These actors can collect data from the sensors and perform appropriate actions as a result of processing such data. To perform the actions at all parts of the region in a timely manner, the actors should be deployed in such a way that they might be able to communicate with each other and cover the whole monitored area. This requires that the actors should be placed carefully prior to network operation in order to maximize the coverage and maintain the inter-actor connectivity. In this paper, we propose a distributed actor deployment algorithm that strives to maximize the coverage of actors without violating the connectivity requirement. The approach applies repelling forces between neighboring actors and from the sensors that sit on the boundaries in order to spread them in the region. The spreading of the nodes is done using a tree of actors which can provide more freedom for the movement of the nodes but at the same time maintain the required connectivity among the nodes. We present two techniques for creation of such an actor tree which are based on local pruning of the actor links and spanning tree of the inter-actor network. The performance of our approach is validated both analytically and experimentally.     

Dispensing and hermetic sealing Rb in a miniature reference cell for integrated atomic clocks

An innovative method for fabricating the reference cell for a Rubidium (Rb) integrated atomic clock is presented. This method uses low-temperature solder sealing technique for producing mini-cells of the size of 14 mm × 10 mm × 3 mm, suitable for Rb miniature atomic clocks. Top and bottom of the cell consists of two glass slides. An LTCC (low-temperature cofired ceramic) module of 2 mm thickness, equipped with a small reservoir for confining the Rb is placed in between the two walls, acting as a spacer, increasing the total volume of the cell. A solder ring joins together the LTCC and the top of the cell. This paper also presents a new technique for handling the Rb, which allows its safe handling and long storage. The alkali metal is stored inside a pool of dodecane, which protects it from oxidation. Pure liquid Rb is then dispensed inside the adjacent reservoir using a glass micropipette; finally, the cell is heated in vacuum, in order to carry out the sealing. The achieved sealing hermeticity was tested, without Rb, by sealing N₂O gas and monitoring its pressure through absorbance measurements using FTIR spectroscopy. Hermeticity was also tested with Rb by integrating a pressure sensor in the LTCC module.     

Throughput behavior of reliable stream protocols in admission controlled bearer networks

This paper conducts an investigation of the application-level behavior of feed-forward rate control mechanisms in broadband networks. Specifically, some potential problems which can have a pronounced impact on performance are identified, by studying the behavior of the Transmission Control Protocol (TCP) overlayed above a generic rate controlled bearer network. Our conclusions, based on some theoretical models in a simulation experiment, demonstrate that under the strict policing approach for rate control, even minor variations in the traffic characteristics or network response can potentially lead to significant degradation in the application-level performance. It is further shown that the adoption of the buffered form of rate control, or traffic shaping, instead of strict policing can alleviate this problem to a significant extent.     

Pressure prediction on a variable-speed pump controlled hydraulic system using structured recurrent neural networks

This paper presents a study to predict the pressures in the cylinder chambers of a variable-speed pump controlled hydraulic system using structured recurrent neural network topologies where the rotational speed of the pumps, the position and the average velocity of the hydraulic actuator are used as their inputs. The paper elaborates the properties of such networks in extended time periods through detailed simulation- and experimental studies where black-box modeling approaches generally fail to yield acceptable performance. As alternative estimation techniques, both linear- and extended Kalman filters are considered in this paper. The estimation properties of the devised network models are comparatively evaluated and their potential application areas are discussed in detail.     

Enhancing data delivery with density controlled clustering in wireless sensor networks

Microsystem Technologies - Wireless sensor networks (WSN) are primarily used for sensing and collecting the information from environment. This information is sent to base station (BS), where, it is...     

Social networks,communication styles,and learning performance in a CSCL community

The aim of this study is to empirically investigate the relationships between communication styles, social networks, and learning performance in a computer-supported collaborative learning (CSCL) community. Using social network analysis (SNA) and longitudinal survey data, we analyzed how 31 distributed learners developed collaborative learning social networks, when they had work together on the design of aerospace systems using online collaboration tools. The results showed that both individual and structural factors (i.e., communication styles and a pre-existing friendship network) significantly affected the way the learners developed collaborative learning social networks. More specifically, learners who possessed high willingness to communicate (WTC) or occupied initially peripheral network positions were more likely to explore new network linkages. We also found that the resultant social network properties significantly influenced learners’ performance to the extent that central actors in the emergent collaborative social network tended to get higher final grades. The study suggests that communication and social networks should be central elements in a distributed learning environment. We also propose that the addition of personality theory (operationalized here as communication styles) to structural analysis (SNA) contributes to an enhanced picture of how distributed learners build their social and intellectual capital in the context of CSCL.     

Passive-space and time view: vector clocks for achieving higherperformance, program correction, and distributed computing

We have noticed two problems with viewing a process as a sequence of events. The first problem is the complete loss of information about potential intra-process concurrency for both sequential and distributed computations, and partial loss of information about potential inter-process concurrency for distributed computations. The second problem is that the resulting reasoning framework does not lend itself to refinement (from sequential computing or a given set of distributed processes) to a preferable set of distributed processes. We argue that it is more natural to view a computation, either distributed or sequential, as a partially ordered set of events. Doing so leads to a view, called passive-space and time view, which we propose. To aid users of the relation “Affects” in developing algorithms, we define vector clocks, that are global logical clocks, so that the relation “Affects”, and hence all potential concurrency, between events can be identified from their timestamps assigned     

On the complexity of solid-state diffusion in highly concentrated alloys and the sluggish diffusion core-effect

Diffusion paths in system V – Nb – W – Al in the highly concentrated region (i.e. as “High Entropy Alloys”) are modelled using DICTRA code with commercial databases and with a new algorithm, based on the Cluster Variation Method (CVM). The DICTRA simulations, which produce realistic diffusion profiles, show that the diffusion path is non-linear. The CVM simulations, with parameters obtained by ab initio calculations, allow to investigate in an incremental way the effects of thermodynamics, of correlations and the role of the intrinsic diffusivities of the components. The results confirm the non-linear character of the diffusion path, and show that it is primarily driven by the complex thermodynamics characteristic of these alloys. When the intrinsic diffusivities are used, it is observed that the slowest diffusing species (in the present case, W) will constrain the whole diffusion process. This is likely the origin of the so called “sluggish diffusion” effect and it is of kinetic origin, bearing no relation with the configurational entropy.     

Resource allocation in computer networks by a regional accessibility model

The problem of resource allocation in computer networks is modeled as a regional accessibility model. These models, which have been used in urban and regional planning, provide a means for planning and evaluating computer networks. We present a straightforward solution procedure which allows for the efficient solution of large models and is well suited to studying the sensitivity of the model to various parameter changes. The model and solution technique are used to study a computer network for the Finnish universities.     

异构蜂窝网络中联合功率控制的终端直通通信资源分配

针对异构蜂窝网络中终端直通（D2D）用户和小蜂窝用户复用宏蜂窝用户资源产生的干扰问题,提出一种联合功率控制的资源分配方案。首先,在满足用户信号干扰噪声比（SINR）和发射功率约束条件的前提下,根据系统干扰模型推导出每个D2D用户和小蜂窝用户复用宏蜂窝用户信道资源时的最优发射功率;其次,将用户的信道选择规划成用户和信道之间的双边匹配问题,采用延迟接受（Gale-Shapley）算法得到一个稳定的匹配解;最后,以所得的匹配解为初始条件,通过交换搜索算法进一步优化分配方案。仿真结果表明,该方案的系统总容量和能量效率分别是最优解的93.62%和92.14%,与随机资源分配方案、无功率控制和交换搜索的分配方案,以及有功率控制无交换搜索的分配方案相比,系统容量平均增幅分别为48.29%、15.97%和4.8%,系统能量效率平均增幅分别为62.72%、44.48%和4.45%。该方案能够达到近似最优的系统总容量,有效提高频率利用率和能量效率。     

Ad Hoc网络基于地理位置的双信道功率退避策略

在地理位置信息的基础上,提出一种双信道功率退避的Ad Hoc网络节能方案。基本思想是控制信道功率由大到小退避,数据信道功率由小到大“退避”。并利用地理位置提供的便利,较好地解决了链路的不对称问题。此外,介绍了功率退避算法,简要分析了成功传输的概率。仿真结果表明,该协议能够满足Ad Hoc网络近距离节点优先通信的设计需求。     

Assigning discounts in a marketing campaign by using reinforcement learning and neural networks

In this work, RL is used to find an optimal policy for a marketing campaign. Data show a complex characterization of state and action spaces. Two approaches are proposed to circumvent this problem. The first approach is based on the self-organizing map (SOM), which is used to aggregate states. The second approach uses a multilayer perceptron (MLP) to carry out a regression of the action-value function. The results indicate that both approaches can improve a targeted marketing campaign. Moreover, the SOM approach allows an intuitive interpretation of the results, and the MLP approach yields robust results with generalization capabilities.