摘要:The existing automated trust negotiation systems suffer from the following drawbacks: the policy languages are coarse-grained, the negotiation strategy cannot generate all trust sequences, and the mechanism to evaluate and select trust sequences is absent. To address the above problems, an attribute-constraint-oriented automated trust negotiation(ACATN)model is proposed. The policy language is refined by using attribute constraint, which can not only effectively protect sensitive services and certificates, but also enhance its flexibility. The global access control policy is used to terminate impossible negotiation in advance so that the negotiation efficiency can be improved. The process of generating trust sequence is described by trust sequence searches tree. Based on the tree, the breadth-first and depth-first searches generate not only a trust sequence quickly, but also all trust sequences. Trust sequences are evaluated via disclosure cost and communication overhead so that the optimal one can be selected. ACATN is illustrated using a typical example.
关键词:access control;automated trust negotiation;negotiation system strategy;attribute constraint
摘要:To address the issue that most of traditional query term proximity(QTP)statistics are lack of effective precision improvement, an analysis method of query terms based on isolated terms filtrating is proposed. The method is based on the fact that the query terms with higher similarity have higher substitutability. Isolated terms in-queries and isolated terms in-documents are extracted from query terms and their instances. The filtrating of isolated terms is earlier than the analyzing of QTP, so that the isolated terms are ignored in the calculating of QTP statistics and that the impact of overemphasis of QTP on precision is reduced. Experimental results on queries with significantly different proximities show that the performance of the QTP with isolated terms filtrating is higher(about 14%)than that of the QTP without isolated terms filtrating is.
关键词:information retrieval;query term proximity;isolated term;term similarity
摘要:The overhead of monitoring that increases with the increasing monitoring tasks is crucial to the effective management and efficient utilization of the cluster computers and to the performance of the monitoring system. Researches are made to reduce the overhead of monitoring by identifying the main characteristics of the application. Main factors of the application are dynamically identified by performing principal component analysis(PCA)on the fly of application execution. The set of the main characteristics is identified through matching up with the comparison characteristic set that is created by 4 categories resource intensive benchmark so that the monitoring workload is reduced. A prototype monitoring system adopting the proposed strategy is implemented. Experimental results show that the monitoring workload is decreased by 20% to 60% in hybrid applications or some resource intensive applications. Large collected data can be effectively refined and processed before inputting it to a monitoring system.
摘要:In mobile ad hoc networks, selfish nodes show non-cooperative behaviors due to discarding packets to save their resources(such as energy), which will seriously affect the network performance. A new cooperation enforcement scheme called CMC is proposed to mitigate this problem. In the CMC, the common neighbor monitoring technique is introduced to help the watchdog monitor all packets transmitting around it so that the system can detect the non-cooperative nodes more quickly. In the routing discovery phase, the route request packets and the route reply packets that contain non-cooperative nodes are dropped by CMC, so that the probability that a well-behaved node uses a bad route for data transmission is reduced. The ns-2 simulation results show that the CMC can improve the throughput of well-behaved nodes by 10%-40% in the presence of 10%-60% non-cooperative nodes.
关键词:mobile ad hoc network;node cooperation;common neighbor monitoring;reputation
摘要:Static race detection techniques consume no extra run-time cost but have lower precision, while dynamic ones have higher precision but consume extra run-time cost due to instrumentation. A new precise and efficient algorithm on incrementally detecting potential data races in Java programs is presented, which is implemented as a race detection pass in the just-in-time(JIT)compiler of the Java virtual machine. The algorithm combines lockset-based and happens-before-relation-based detection. Then the algorithm does an intra-method analysis on each method compiled by JIT in turn, and collects summaries independent of the context. The context-sensitive inter-thread analysis is proposed based on the method summaries, to compute incremental race information. The resulting information is output in time. Experimental results show that the algorithm has no instrumentation cost and unlimited program scale, and that the algorithm has the similar precision as O'Callahan, et al's algorithm on dynamic race detection, and consumes only 2%-4% of the total compilation time.
摘要:According to the communication capability of current sensor nodes, a novel scheme(REDD)is proposed for data delivery in large-scale wireless sensor networks. In REDD, the clustering network architecture is formed by periodically selecting cluster heads among sensor nodes through dynamic competition. The intra-cluster communication is based on TDMA scheduling, in which a member node transmits its sensory data through an independent intra-cluster channel to its cluster head in the time slot assigned by the cluster head. The inter-cluster communication is based on CSMA competition, in which a connected overlay network is constructed among all cluster heads to forward the aggregated data through multi-hops to the sink. Simulation results show that, compared with the existing work, the REDD can achieve a better clustering performance with lower costs, and effectively reduce the end-to-end delay for reliable data transmission.
摘要:Aiming at the locating problem in large scale wireless sensor networks, a local SDP algorithm(LSDP)is proposed. A large scale wireless sensor network is split into some local networks according to a node importance index of each node, and a relative map in each local network is then obtained with SDP. All relative maps are patched into a global relative map. After matching with reference nodes, the global relative map is transformed into a global absolute map, and then all nodes in the network are located. Simulation results show that the location error is as little as 98% of that in ideal communications when WSN is regularly deployed. The location error is reduced by 90% if WSN is randomly deployed.
摘要:A novel distributed localization method based on radio interference(DRIPS)is proposed to reduce the complexity of existing localization methods for wireless sensor networks.The DRIPS utilizes two anchors to create an interference signal. Based on the Doppler effect produced by the movement of one transmitter, each sensor in a network can obtain the angle information with the mobile anchor through the Doppler shift measurement of locally received signal strength indication signal. Finally, each sensor utilizes multi-sets of the angle information and the location of the mobile anchor to estimate its optimal geographical location independently. The DRIPS only relies on radio transceivers without the use of extra measurement equipments, and the sensor node conducts localization in a completely distributed manner without information exchange or centralized processing. Simulation results show that the proposed method is effective and that the average localization error in a network will be less than 1 m when the measurement accuracy of Doppler shift is 1 Hz.
摘要:Focusing on the problem that the bit error rate performance of the space-time block codes decreases in frequency-selective fading channels, a time-reversal coordinate interleaved orthogonal coding design(TR-CIOD)is proposed. Data symbols are transmitted in designed format and some of the symbols are time-reversed. It is the orthogonality of the equivalent channel matrix and the multiplication property of the circulant matrix that the influence on performance caused by intersymbol interference is greatly reduced after matching filters at receivers. TR-CIOD can achieve multi-antenna diversity gain as well as multi-path diversity gain. Simulation results with 10-5 bit error rate and 2-ray channels and comparisons with the coordinate interleaved orthogonal design in frequency flat channel show that TR-CIOD performs obviously better, and achieves about 5.5 dB SNR gain.
摘要:Focusing on the high computational complexity of partner selection methods in cooperative systems, two improved partner selection methods are proposed. These are the best channel matching strategy and the non-reciprocal best channel matching strategy. The user with the worst link is firstly taken into account in the best channel matching strategy. Then the user with the best inter-user channel quality is selected by the method as his partner. The two users cooperate to relay information for each other, and unnecessary cooperative process is avoided by setting up two special thresholds of signal-to-noise. The non-reciprocal best channel matching strategy starts from the user with the worst link and then selects the user with the best inter-user channel as a partner for each user. Simulation results show that both the partner selection strategies can reduce O(N2)computation complexity, meanwhile keep the same energy gain as the existing worst-link-first algorithm.
摘要:Focusing on the problem that the extrinsic information transfer(EXIT)functions of the iterative processing unit in existing methods are of low accuracy, an accurate method named optimal histogram(OptHIST)is proposed. The OptHIST algorithm is implemented in the following steps. The log-likelihood ratio data are classified according to the value(+1 or-1)of the info bits, and the optimal histogram is then used to estimate the probability distribution function of the log-likelihood ratio. The optimal probability distribution function is obtained in the sense of the integrated mean squared error.Then the exact EXIT function is obtained through integrating the probability distribution function of log-likelihood ratio. Compared with the existing methods, the proposed OptHIST method is more robust than the histogram method and more applicable than the direct average method. Simulation results show that the OptHIST method can reduce the error by about 7% to 15%, compared with the histogram method, for the processing unit which adopts rigid a-post probability algorithm.
关键词:iterative processing unit;extrinsic information transfer function;optimal histogram
摘要:According to the impacts of unit maintenance on the system operation costs, a novel unit maintenance scheduling(UMS)model is proposed in a competitive environment. Based on the system reliability assessment, the minimized total system costs over the entire horizon is taken as the objective, and the constraints of the system loss of load probability(LOLP), power flow and maintenance outage are considered completely. Compared with the existing models, the proposed scheme takes the influence of bilateral contracts and energy market as well as random failures of generating units and transmission lines into consideration. Additionally, the system reliability index is converted into the economic index as a part of the objective,and evaluated by Monte Carlo approach. The simulation results illustrate that the present model ensures the system LOLP to be kept under the pre-set tolerance value with a steady outage planning cost.
摘要:A PSOGM(1,1)model is proposed with internal variable parameters identification to improve the precision of the traditional GM(1,1)model for electricity price forecasting in power markets. The intensional expressions that describe the nonlinear relations between internal variable parameters(such as developing coefficients, the grey inputs, the background weight parameters, and the boundary-values)and forecasting values are deduced. Then the particle swarm optimization algorithm(PSO)is adopted to identify the internal parameters. The optimal solution to the model is obtained and the PSOGM(1,1)model is generated. Comparisons with the traditional GM(1,1)model show that the PSOGM(1,1)model provides faster convergence rate and better prediction precision. Numerical results on the historical data of NORDPOOL power market show that the average precision of the PSOGM(1,1)model is 94% for short-term price forecasting, and is 1% to 3% higher than the traditional GM(1,1)model and other typical improved GM(1,1)models.
摘要:To improve the overall performance of the three-level active power filter(APF), the closed-loop control system of this APF is investigated, and it is found that tracking command harmonic current is one of the keys to designing this APF. The current state feedback decoupling PI control is analyzed in detail and the limitation of the conventional PI conditioner in the traditional strategies, a novel PI control method is proposed, where the predictive output current value is obtained by the state predictor, and the delay of one sampling period is remedied in this digital control system by the state predictor. The predictive command harmonic current value is obtained by the repetitive predictor synchronously with higher dynamic convergence rate and little static misadjustment. The experiment results demonstrate the steady filtering capability.
关键词:active power filter;three-level inverter;PI control;predictor
摘要:To solve the uncertainty of parameter variations and load disturbance, a nonlinear decoupling control based on disturbance suppression is presented for the permanent magnet synchronous motors drive system,where the adaptive control is combined with nonlinear decoupling control, and the parameter variations and load disturbance are regarded as the disturbance input. The state feedback adaptation law and parameter adaptation law are designed with Lyapunov theory to ensure the system convergence and the asymptotic tracing based on L2 transient characteristics. The simulation and experiments show the greatly improved dynamic performance, robustness and anti-disturbance capacity of the system.
关键词:nonlinear control;permanent magnet synchronous motor;decoupling control;adaptive control;L2 gain control
摘要:A technique about spatial cold shield system with phase-change refrigeration by the cryogens is presented to lower the surface temperature of a spacecraft under 100 K, furthermore to lower the infrared radiation on the spacecraft under 0.5 W/m2 in 30 minutes, then a 3-D layered reservoir model is established for the simulation of heat transfer. In this simulation process, the empirical equations are substituted to boundary conditions of stable heat-transfer relations to simulate the surface temperature and heat flux changing with the liquid level. Then an experimental unit model is constructed for process analysis. The simulation and experiment results show that the experimental temperature gradients are similar to the simulation results before the exhaustion of liquid nitrogen when liquid level of layered reservoir is less than 100 mm; surface temperature of the system meets the requirement of infrared radiation in 15 min on the ground and 30 min in the space; solar radiation and capillary material influence the uniformity of surface temperature and holding time of liquid nitrogen to some degree.
摘要:A multi-round scheduling algorithm, data-collection multi-round(DCMR), is presented to minimize the makespan of divisible workloads in parallel computing. A three-stage model is proposed and takes communication latency and computation start-up time into consideration. The algorithm provides a method to generate a near-optimal number of scheduling rounds. Close-form equations are given through analyzing a specific time sequence of load distribution, and then the bisection method, combined with back-forward adjustment, is used to get an asymptotically optimal number of scheduling rounds, which make the computation time overlap the communication time as much as possible and reduce the makespan. Simulation results show that the algorithm can find a near-optimal number of scheduling rounds under different network parameters. Compared with the classical algorithms such as FIFO and LIFO, the DCMR has higher adaptability. When the computation time dominates the communication time, the algorithm can keep the makespan at a rather low level which is about 1.1 times of the ideal time.
摘要:Glycerol monostearate(GMS)and glycerol monolaurate(GML)were investigated by viscometer and microscope. The results show that foams of GMS gets more steady than GML at the same concentration, the survival time of GMS/4%(mass fraction)foams persists over 20 days. The concentration of surfactant is the main factor for the survival time. Once the concentration exceeds the critical micelle concentration(CMC), a great deal of lamellar liquid crystal(LLC)displaying the cross figure is discovered in the GMS solution, conversely, the extremely small amount in the GML, so the LLC may increase the foams stability. Furthermore, the same experiment was conducted to exhibit the identical conclusion by sodium dodecyl benzene sulfonate(SDBS)and α-olefin sulfonate(AOS). Three important factors enhance the stability of foams via slowing down the drainage speed of liquid films, increasing the Gibbs-Marangoni mechanism, and equalizing the bubble size.