语言型群组决策的专家权重微调整方案

针对语言型偏好的群组决策问题,提出一种利用专家权重微调整进行决策优选的方法.首先定义了转换函数和逆转换函数.将语言型评价通过转换函数转变为数值型,然后运用专家权重的微调整方案为专家赋权,并依据此权重利用加权和进行方案排序,最后用逆转换函数将结果转换为语言型结论.

Selective harmonic elimination method for wide range of modulation indexes in multilevel inverters using ICA

Selective harmonic elimination(SHE) in multilevel inverters is an intricate optimization problem that involves a set of nonlinear transcendental equations which have multiple local minima. A new advanced objective function with proper weighting is proposed and also its efficiency is compared with the objective function which is more similar to the proposed one. To enhance the ability of the SHE in eliminating high number of selected harmonics, at each level of the output voltage, one slot is created. The SHE problem is solved by imperialist competitive algorithm(ICA). The conventional SHE methods cannot eliminate the selected harmonics and satisfy the fundamental component in some ranges of modulation indexes. So, to surmount the SHE defect, a DC-DC converter is applied. Theoretical results are substantiated by simulations and experimental results for a 9-level multilevel inverter. The obtained results illustrate that the proposed method successfully minimizes a large number of identified harmonics which consequences very low total harmonic distortion of output voltage.

安全高效的可撤销指纹模板构造

指纹特征加密是生物特征识别技术中最为活跃的研究领域之一。然而指纹特征是人体固有的特征,具有唯一性与终身不变性,一旦识别系统中的用户指纹模板丢失,将是永久性的。为了避免用户指纹模板丢失,基于巴特沃斯低通滤波器原型,设计了一组不可逆转换函数,将用户原始指纹模板转换为可撤销指纹模板,不同识别系统中存储的可撤销指纹模板具有互异性与不可逆性,进而有效规避原始指纹模板的不变性与公开性所带来的安全隐患。匹配精度、不可逆性、转换速度等方面的实验分析表明,所构造可撤销指纹模板方案在认证效率、安全性、计算效率等方面具有良好性能。