ID 原文 译文
38516 为规避参数估计过程中模型函数导数连续性和存在性的限制,应用一个自适应变化的实值遗传算法进行计算。 To avoid in the process of the parameter estimation model function derivative continuity and the existence of the limit, using an adaptive change of real-valued genetic algorithm to calculate.
38517 最后通过一个真实软件失效数据集对模型参数进行计算,利用不同的评价准则将所提模型和现有模型进行比较,说明所提模型的优越性和准确性。 Finally, a real software failure data set of model parameters is calculated, by using different evaluation criteria we compare the proposed model and existing model, and illustrate the superiority of the proposed model and the accuracy.
38518 为求得类似仿真函数的黑箱函数优化问题的全部局部极值点,提出了一种基于适应值曲面分析的新算法。 To obtain similar simulation function of whole local extremum points, black-box function optimization problem put forward a new algorithm based on adaptive value surface analysis.
38519 首先,对适应值距离相关系数(fitness distance correlation,FDC)进行了改进,探讨了改进FDC指标与适应值曲面崎岖度的对应关系。 First of all, to adapt to the value from the correlation coefficient (fitness short the correlation, FDC) is improved, and improved FDC index is discussed to adapt to the value of surface rough degree of corresponding relation.
38520 在此基础上,设计了基于FDC的重复对分区域法(FDC based repeated split region,FRSR),对可行域依据崎岖度进行分解,得到满足崎岖度要求的若干子区间,并在这些子区间上依据FDC指标设置初始点,然后利用模式搜索算法进行寻优。 On this basis, the design of a repeat of subregional method based on FDC (FDC -based repeated the split region, FRSR), the feasible region according to the decomposition of the rough degree, get some subinterval, satisfies the requirement of the rough degree on the subinterval FDC indexes set initial point, and then using pattern search algorithm for optimization.
38521 通过对比FRSR法与传统的均匀分配初始点法以及其他现有方法,验证了FRSR法能够以较少的初始点得到全部局部极值,在速度上和解的质量上都优于传统方法。 FRSR method by comparison with the traditional uniform distribution of the initial point method and other existing methods, verified the FRSR method with less initial points got all the local extremum, reconciliation in pace, quality is better than traditional methods.
38522 针对具有不确定性的并发故障诊断问题,提出基于属性权重和权衡分析的置信规则库(belief rule base,BRB)诊断方法。 In view of the uncertainty of concurrent fault diagnosis problem, put forward is based on the analysis of the attribute weights and weigh the belief rule base (rule base are, BRB) diagnostic method.
38523 以属性权重大小来表示属性与特定故障模式之间的相关性,设计能够反映故障模型约束的基于差分进化的优化算法,通过相邻故障模式置信度与预设阈值的权衡分析完成对并发故障诊断。 To attribute weights to represent the attribute and the correlation between specific failure mode, the design can reflect the fault model constrained optimization algorithm based on differential evolution, through the adjacent fault mode confidence and preset threshold tradeoff analysis of concurrent fault diagnosis.
38524 该方法仅需构造单个置信规则库来有效处理各种不确定性信息,与已有研究方法相比极大地降低了建模复杂度。 This method is only a single belief rule base were constructed to deal effectively with the various uncertainty information, compared with the existing research methods greatly reduce the complexity of modeling.
38525 诊断结果不仅能得到故障的并发情况,还可分辨故障的主次关系,并且建模和推理过程开放,可解释性强。 Not only can get the fault diagnosis of concurrent situation, also can distinguish the progression of failure, and is open to modeling and reasoning process, can be interpreted.