Sep 2006, Volume 1 Issue 3

  • Select all
  • QIANG Lei, XIAO Tian-yuan
    This paper presents a new distributed Bayesian optimization algorithm (BOA) to overcome the efficiency problem when solving NP scheduling problems. The proposed approach integrates BOA into the co-evolutionary schema, which builds up a concurrent computing environment. A new search strategy is also introduced for local optimization process. It integrates the reinforcement learning (RL) mechanism into the BOA search processes, and then uses the mixed probability information from BOA (post-probability) and RL (pre-probability) to enhance the cooperation between different local controllers, which improves the optimization ability of the algorithm. The experiment shows that the new algorithm does better in both optimization (2.2 %) and convergence (11.7 %), compared with classic BOA.
  • DAI Jian-hua, PAN Yun-he
    Rough set axiomatization is one aspect of rough set study to characterize rough set theory using dependable and minimal axiom groups. Thus, rough set theory can be studied by logic and axiom system methods. The classical rough set theory is based on equivalence relation, but the rough set theory based on similarity relation has wide applications in the real world. To characterize similarity-based rough set theory, an axiom group named S, consisting of 3 axioms, is proposed. The reliability of the axiom group, which shows that characterizing of rough set theory based on similarity relation is rational, is proved. Simultaneously, the minimization of the axiom group, which requests that each axiom is an equation and independent, is proved. The axiom group is helpful to research rough set theory by logic and axiom system methods.
  • LIU Yu-sheng, CHEN Jiang, LI Xing-yuan
    Many physical systems such as biochemical processes and machines with friction are of nonlinearly parameterized systems with uncertainties. How to control such systems effectively is one of the most challenging problems. This paper presents a robust adaptive controller for a significant class of nonlinearly parameterized systems. The controller can be used in cases where there exist parameter and nonlinear uncertainties, unmodeled dynamics and unknown bounded disturbances. The design of the controller is based on the control Lyapunov function method. A dynamic signal is introduced and adaptive nonlinear damping terms are used to restrain the effects of unmodeled dynamics, nonlinear uncertainties and unknown bounded disturbances. The backstepping procedure is employed to overcome the complexity in the design. With the proposed method, the estimation of the unknown parameters of the system is not required and there is only one adaptive parameter no matter how high the order of the system is and how many unknown parameters there are. It is proved theoretically that the proposed robust adaptive control scheme guarantees the stability of nonlinearly parameterized system. Furthermore, all the states approach the equilibrium in arbitrary precision by choosing some design constants appropriately. Simulation results illustrate the effectiveness of the proposed robust adaptive controller.
  • WANG Ding-wei, ZHANG Guo-xiang
    The location of rescue centers is a key problem in optimal resource allocation and logistics in emergency response. We propose a mathematical model for rescue center location with the considerations of emergency occurrence probability, catastrophe diffusion function and rescue function. Because the catastrophe diffusion and rescue functions are both nonlinear and time-variable, it cannot be solved by common mathematical programming methods. We develop a heuristic embedded genetic algorithm for the special model solution. The computation based on a large number of examples with practical data has shown us satisfactory results.
  • KUANG Ping, ZHU Qing-xin, WANG Ming-wen, CHEN Xu-dong, QING Li
    To improve the computational performance of the fuzzy C-means (FCM) algorithm used in dataset clustering with large numbers, the concepts of the equivalent samples and the weighting samples based on eigenvalue distribution of the samples in the feature space were introduced and a novel fast cluster algorithm named weighted fuzzy C-means (WFCM) algorithm was put forward, which came from the traditional FCM algorithm. It was proved that the cluster results were equivalent in dataset with two different cluster algorithms: WFCM and FCM. Furthermore, the WFCM algorithm had better computational performance than the ordinary FCM algorithm. The experiment of the gray image segmentation showed that the WFCM algorithm is a fast and effective cluster algorithm.
  • ZHANG Feng, FAN Xiao-zhong
    The resolution of overlapping ambiguity strings (OAS) is studied based on the maximum entropy model. There are two model outputs, where either the first two characters form a word or the last two characters form a word. The features of the model include one word in context of OAS, the current OAS and word probability relation of two kinds of segmentation results. OAS in training text is found by the combination of the FMM and BMM segmentation method. After feature tagging they are used to train the maximum entropy model. The People Daily corpus of January 1998 is used in training and testing. Experimental results show a closed test precision of 98.64 % and an open test precision of 95.01 %. The open test precision is 3.76 % better compared with that of the precision of common word probability method.
  • CHEN Wei, YANG Yi-xian, NIU Xin-xin
    Considering connections of characteristics, this paper is aimed at the construction of optimized Boolean functions. A new method based on the Bent function, discrete Walsh spectrum and characteristics matrices are presented by concatenating, breaking, and revising output sequences conditionally. This new construction can be used to construct different kinds of functions satisfying different design criteria.
  • HONG Fan, DUAN Su-juan, LI Cheng-bing
    Trust management was discussed according to the security requirements in distributed environment. A graph-based delegable model and its formal description were given. Permissions propagating to control, conflict and cyclic in delegable model were analyzed and resolved. Finally, the scheme of determining access request of subject in distributed environment and future work was proposed.
  • JIANG Zheng-tao, SUN Xi, WANG Yu-min, TIAN Lei
    Properties of third-order recurrence sequences were investigated and a new variant of the GH public-key cryptosystem, which was further improved to be a probabilistic public-key cryptosystem, was proposed. Then security analysis of the proposed scheme was provided and it was proved that the one-wayness of the proposed scheme is equivalent to partial discrete logarithm and its semantic security is equivalent to decisional Diffie-Hellman problem in ring extension. Finally, efficiency analysis of the proposed scheme was provided, and that these two encryption schemes need to transfer 2log N and 4log N bits data respectively.
  • SUN Chang-hua, LIU Bin, LI Wen-jie
    A general model is made to analyze switching throughput of traffic manager in core routers. By designing a real traffic manager that uses the OC-48c interface, the whole system is analyzed and it is pointed out that at least four HSSLs should be employed per CSIX interface when using Vitesse s GigaStream switch chipset. Meanwhile, at the CSIX interface, the CFrame should be constructed according to the actual size of the last cell of each IP packet. The above principles can guarantee forwarding of IP packets at line rate. A general relationship between throughput and buffering scheme of IP packets in the external memory is given, which is useful in the design of switch fabric in core routers.
  • LIN Zhi-ting, ZHAO Bao-hua, QU Yu-gui, ZHAI Yu-jia
    An effective method for extending wireless sensor network lifetime is presented for redundancy nodes. Because most of the existing coverage protocols are based on the circular sensing model, this paper discusses how to combine consideration of coverage and connectivity maintenance in a single effective scheduling when the sensing model is not circular. The least redundancy coverage strategy (LRCS) strategy takes the smallest redundant coverage as the criterion whose goal is to maximize the lifetime of the sensor network. The sensor nodes are divided into five statuses: the judgment status, the waiting status, the activity status, the dormancy status and the exhausted status. Simulations show that LRCS outperforms the coverage configuration protocol and is suitable for various situations. It effectively reduces the number of active nodes and extends the lifetime of the sensor network.
  • LI Zan, CHANG Yi-lin, WANG Yu-min, CAI Jue-ping
    A novel family of frequency-hopping (FH) sequences based on iterated block cipher is proposed for frequencyhopping multiple-access (FHMA) communications. The design offers a class of nonlinear FH codes with high security, large linear span and a uniform spread over the entire frequency bandwidth. Moreover, FH sequences among the family are independent from each other and they perform as well as random patterns in terms of multiple access interference in anti-jamming applications. With the performance of packet error and throughput for FHMA network being derived in theory, many numerical results of the 3DES sequences are presented, comparing with those of shift register sequences and chaotic FH sequences. Efficiently implemented in field programmable gate arrays (FPGA), the generator prototype of the proposed sequence has been realized and incorporated into fast FH radio.
  • GUO Jie, LIU Yun, YE Zhi-hui, SONG Tie-cheng, SHEN Lian-feng
    This paper studies an algorithm about separating spectrally overlapping signals using the cyclostationary properties of signals. On the basis of direct sequence spread system (DSSS), frequency shift filter is added into the receiver of the communication system. Although the structure of frequency shift filter is more complicated than the time-domain filter, it uses both time correlations and frequency spectrum correlations so it can achieve better performances on separating the overlapping signals. After the analysis of cyclostationary characteristic and frequency spectrum correlation, the structure of the frequency shift filter can be gained. Then, a self-adaptive algorithm is utilized for the purpose of achieving optimum multidimensional tap weights of frequency shift components. The simulation results indicate that this method can efficiently separate overlapping signals, and its error rate is lower than the time-domain filter or DSSS system by two orders of magnitude on the condition that high-power interference is added into the system.
  • YUE Guang-rong, LI Shao-qian, LI Chuan
    This paper is focuses on the unbalanced bit error rate (BER) for different transmitted data symbols of ultra wideband (UWB) impulse radio time-hopping pulse-position-modulation (TH-PPM) in a deterministic multi-path environment for unreasonable parameter selection. Two solutions are presented here. The first is that the decision threshold of RAKE is dynamic for different channel environments; the other is that we can improve the traditional TH-PPM modulation, which is, encoding the transmitted data symbol with balance code. It is shown by theoretical analysis and computer simulation that these two methods can solve the unbalanced BER of traditional TH-PPM. Compared with the dynamic threshold method, the balance encoding scheme can be implemented more easily, and is more robust to the channel time variant characteristics, the channel estimation of RAKE receiver and the combination techniques.
  • WANG Yi-ming, ZHU Hong-bo
    In this paper, a new adaptive equalization scheme aiming at the intersymbol and interuser interferences in the multiuser ultra-wide band (UWB) channels is presented specially for direct-sequence UWB/time-hopping UWB (DS-UWB/TH-UWB) receivers. Its computational complexity and convergence rate are compared with the conventional algorithms. The simulation results show that the computational complexity of this scheme is less than employing the RLS method alone and the convergence rate of this scheme is farther off employing LMS method alone.
  • ZHAO Hai, SU Wei-ji, XU Ye, ZHANG Xin
    Separation degree is a standard measure for complex network research. Whatever its scale or its increase makes the Internet take on a complex network character. Because of the development of complex network theory and the continuous evolution of the Internet, it is a key problem that uses complex network theory to research the Internet nowadays. In this paper, the Internet separation degree is put forward. The time series stochastic process model of the Internet separation degree is established. According to actual data, the Internet separation degree time sensitivity model (ISDTSM) is established and the effect of time sensitivity of the Internet separation degree to the Internet IP level transmission is computed. Finally the Internet separation and IP transmission during 2008 Beijing Olympic Games were forecasted by using the model.
  • LIU Jun-min, ZHANG Chen, LIU Shi
    A new positioning method in mobile networks is presented. Based on the data fusion technology, it processes multi-layer information fusion for the location estimates achieved by the Chan algorithm, which increases mobile positioning accuracy effectively by only using measured difference of arriving (TDOA) signals. The method is simple and practical, especially when the location estimates are corrupted by the non-line-of-sight (NLOS) error. It not only has high positioning accuracy, but also reduces the location failure probability. Results from computer simulation show that the proposed method is effective in various environments.
  • WANG Zhen-li, ZHANG Xiong-wei, YANG Ji-bin, CHEN Gong
    A new fast adaptive filtering algorithm was presented by using the correlations between the signal s former and latter sampling times. The proof of the new algorithm was also presented, which showed that its optimal weight vector was the solution of generalized Wiener equation. The new algorithm was of simple structure, fast convergence, and less stable maladjustment. It can handle many signals including both uncorrelated signal and strong correlation signal. However, its computational complexity was comparable to that of the normalized least-mean-square (NLMS) algorithm. Simulation results show that for uncorrelated signals, the stable maladjustment of the proposed algorithm is less than that of the VS-NLMS algorithm, and its convergence is comparable to that of the algorithm proposed in references but faster than that of L.E-LMS algorithm. For strong correlation signal, its performance is superior to those of the NLMS algorithm and DCR-LMS algorithm.
  • KE Pin-hui, KE Pin-hui, CHANG Zu-ling, CHANG Zu-ling, WEN Qiao-yan, WEN Qiao-yan
    Bent functions in trace forms play an important role in the constructions of generalized binary Bent sequences. Trace representation of some degree two Bent functions are presented in this paper. A sufficient and necessary condition is derived to determine whether the sum of the combinations of Gold functions, tr1n (x2i+1 ), 1≤in - 1, over finite fields F2n (n be even) in addition to another term tr1n/2 (x2n/2+1 ) is a Bent function. Similar to the result presented by Khoo et al., the condition can be verified by polynominal greatest common divisor (GCD) computation.A similar result also holds in the case Fpn ( n be even, p be odd prime). Using the constructed Bent functions and Niho type Bent functions given by Dobbertin et al., many new generalized binary Bent sequences are obtained.
  • LIU Shao-dong, JIAO Yong-chang, ZHANG Fu-shun
    In this paper the average power pattern of the side-fed offset Cassegrain (SFOC) dual reflector antenna is analyzed, and the effect of the random surface error on radiation characteristics of the antenna is introduced. Here, the random surface error is defined as the error of the standard reflector in its normal direction and the errors in a small zone of the reflector are considered as equal. We also assume that the phase error on the aperture led by the random surface error obeys a Gaussian distribution with zero mean, under which the expression of the average power pattern is deduced. Finally, the data related to the radiation characteristics of the antenna are calculated and the corresponding curves are presented. The obtained results can be used for the user to determine the manufacturing accuracy of the reflector of the SFOC antennas.
  • XU Yue-tong, FU Jian-zhong, CHEN Zi-chen
    Thrust ripple such as end force, slot force and normal force are key factors that affect the properties of permanent magnet linear synchronous motors (PMLSM). According to different mechanics and analytical models, end force resulting from open magnetic circuit of PMLSM was greatly decreased by optimizing the length of the PMLSM mover. Slot force caused by slot effect was greatly reduced by using fraction slot structure, and normal force was calculated through the finite element method (FEM). After thrust ripple was calculated, its uniform formula was obtained through Fourier series nonlinear regression. An experimental system was set up to measure thrust ripple, and experiment results demonstrated that experimental force ripple was quite in line with that calculated by the fitting formula. The optimal theory and analysis method is effective, and the obtained formula can be utilized to compensate thrust ripple in practical applications and improve the motion performance of PMLSM.
  • HUANG Ping-lin, HU Qian-sheng, YU Li, HUANG Yun-kai
    Based on the configuration of deep slot concentrated coils, an analytical model is developed for predicting the armature reaction field produced by the 3-phase stator windings of permanent magnet brushless DC motors with concentrated coils by using the image method and the analytical functions of the armature reaction and winding inductances are proposed accounting for the influence of stator slotting. This approach is different from the method of equivalent distributed current sheet and more suitable for electric machines, which have concentrated coils and deeper slots. Under different control mode, the different analytical functions are presented. This will be helpful when further analyzing the performance of the motor. The results agree with the experiment very well.
  • SUN Xue-ming, ZHANG Hui-jian, ZUO Meng, GU Wan-yi, XU Da-xiong
    Dense wavelength division multiplexing (DWDM) system is the ultimate selection as an optical communication system because of its high speeds and capacities. However, the fiber nonlinear effects and polarization mode dispersion severely limit the performance of the system when signal propagates at 40 Gbit/s in a single channel. The coupled nonlinear Schröinger equations of a single channel in DWDM, which are all considered factors of group velocity dispersion (GVD), self phase modulation (SPM), cross phase modulation (XPM), four wave mixing (FWM) and polarization mode dispersion (PMD), are derived, while their number results are obtained with extended split-step Fourier method. Finally, to analyze the impacts of the fiber nonlinear effects and PMD on the optical communication system, the simulated results of an 8×40 Gbit/s DWDM system are discussed under different conditions respectively.