交叉熵重要抽样法基于KL距离(Kullback-Leibler divergence)实现了重要抽样概率密度函数(importance sampling probability density function,IS-PDF)参数的有效估计,显著提升了电网可靠性的蒙特卡洛仿真速度,但KL距离仅是更广义的f散...交叉熵重要抽样法基于KL距离(Kullback-Leibler divergence)实现了重要抽样概率密度函数(importance sampling probability density function,IS-PDF)参数的有效估计,显著提升了电网可靠性的蒙特卡洛仿真速度,但KL距离仅是更广义的f散度家族的一种距离形式,虽然有效但非唯一。为探讨f散度在重要抽样法中的实现方式,该文提出最优f散度重要抽样法,通过最小化零方差概率密度函数和IS-PDF之间的f散度,推导典型距离测度下IS-PDF的参数统一迭代更新表达式。因不同距离测度对应不同IS-PDF参数和重要抽样效率,为此,基于可靠性指标测试函数与似然比函数之积的方差最小化准则,提出f散度下的最优距离形式迭代优化确定方法。对IEEE-RTS79和IEEE-RTS96系统进行可靠性评估,结果表明该文方法能有效提高仿真效率,实现重要抽样法效率提升潜力的有效挖掘,对于大电网可靠性的快速评估具有重要的工程实用价值。展开更多
Channel parameters estimation in an orthogonal for the receiver station is a multi-dimensional （MD） frequency division multiple access （OFDMA） system optimization problem, because every user node has a separate lo...Channel parameters estimation in an orthogonal for the receiver station is a multi-dimensional （MD） frequency division multiple access （OFDMA） system optimization problem, because every user node has a separate local oscillator and every transmitter to receiver link has individual carrier frequency offset （CFO） and channel impulse response （CIR） parameters. In order to reduce the computational complexity for MD optimization, a time domain CFOs and CIRs estimation algorithm over the OFDMA based wireless multimedia sensor networks （WMSN） is proposed in this paper. In this algorithm, the receiver station can decouple the signal from every node by correlation based on specially designed training sequences, so that the MD optimization problem is simplified to an 1-D optimal problem. It is proved that the multiple CFOs can be identified from the correlation result using the phase shift of the consecutive training se- quences. Based on the CFOs estimation result, the CIRs can then he estimated according to the minimum mean square error （MMSE） criterion. The theoretic analysis and simulation results show that the proposed algorithm can effectively decouple the signal from different user nodes and the bit error rate （BER） per- formance curves are close to the ideal estimation when the user number is not large.展开更多
文摘交叉熵重要抽样法基于KL距离(Kullback-Leibler divergence)实现了重要抽样概率密度函数(importance sampling probability density function,IS-PDF)参数的有效估计,显著提升了电网可靠性的蒙特卡洛仿真速度,但KL距离仅是更广义的f散度家族的一种距离形式,虽然有效但非唯一。为探讨f散度在重要抽样法中的实现方式,该文提出最优f散度重要抽样法,通过最小化零方差概率密度函数和IS-PDF之间的f散度,推导典型距离测度下IS-PDF的参数统一迭代更新表达式。因不同距离测度对应不同IS-PDF参数和重要抽样效率,为此,基于可靠性指标测试函数与似然比函数之积的方差最小化准则,提出f散度下的最优距离形式迭代优化确定方法。对IEEE-RTS79和IEEE-RTS96系统进行可靠性评估,结果表明该文方法能有效提高仿真效率,实现重要抽样法效率提升潜力的有效挖掘,对于大电网可靠性的快速评估具有重要的工程实用价值。
基金supported by the National High Technology Research and Development Programme of China(No.2006AA01Z216)
文摘Channel parameters estimation in an orthogonal for the receiver station is a multi-dimensional （MD） frequency division multiple access （OFDMA） system optimization problem, because every user node has a separate local oscillator and every transmitter to receiver link has individual carrier frequency offset （CFO） and channel impulse response （CIR） parameters. In order to reduce the computational complexity for MD optimization, a time domain CFOs and CIRs estimation algorithm over the OFDMA based wireless multimedia sensor networks （WMSN） is proposed in this paper. In this algorithm, the receiver station can decouple the signal from every node by correlation based on specially designed training sequences, so that the MD optimization problem is simplified to an 1-D optimal problem. It is proved that the multiple CFOs can be identified from the correlation result using the phase shift of the consecutive training se- quences. Based on the CFOs estimation result, the CIRs can then he estimated according to the minimum mean square error （MMSE） criterion. The theoretic analysis and simulation results show that the proposed algorithm can effectively decouple the signal from different user nodes and the bit error rate （BER） per- formance curves are close to the ideal estimation when the user number is not large.
