论文部分内容阅读
在统计过程控制中,为了利用专家信息来减少不确定性,并在控制图模型中将不同属性的专家先验信息结合起来综合利用,提出了一种新的广义标准灰数概念,将不同属性的专家信息结合在同一个空间上,用统一架构进行表征,并提出了新的运算法则来计算这些多源异构的专家灰信息.结合经典贝叶斯理论,在灰数据背景下对统计质量控制图模型进行参数估计,并利用累积样本信息对参数进行迭代优化,使灰色区域不断收敛,降低不确定性.实例分析结果表明:这种灰贝叶斯迭代优化模型可以在小样本贫信息的情况下减少监测数据的异常波动,更准确地利用专家信息进行预警,并在样本累积过程中逐步偏重于实际数据,得到符合新样本信息的参数.
In the process of statistical process control, in order to reduce the uncertainty by using the expert information and to integrate the priori information of experts with different attributes in the control chart model, a new concept of generalized standard gray number is proposed, Expert information is combined in the same space and characterized by a unified architecture and a new algorithm is proposed to calculate the gray information of these heterogeneous experts.Combined with the classic Bayesian theory, The control chart model is used to estimate the parameters and iteratively optimize the parameters by using the accumulated sample information to keep the gray area convergent and reduce the uncertainty.Example analysis results show that this gray Bayesian iterative optimization model can optimize the parameters in small samples with poor information Reduce the abnormal fluctuation of the monitoring data, and make more accurate use of the expert information to make early warning. In the process of sample accumulation, the data will be more emphasis on the actual data and the parameters that accord with the new sample information will be obtained.