煤矿基层管理者在疲劳状态下的风险感知水平与决策水平快速监测，是煤矿管理中的难题。目前，基于心理量表测量范式，监测与判别煤矿基层管理者风险感知水平与决策水平的方法，具有主观性强、数据获取难度大、测量结果有延迟性等局限性，成为制约监测和识别煤矿基层管理者疲劳风险感知与决策水平的瓶颈。本研究以煤矿基层管理者为研究对象，以不同难度的决策失误案例为刺激条件，以疲劳与正常状态为对比组，采用理论分析和事件相关电位脑电实验相结合的方法，开展疲劳状态与正常状态的煤矿基层管理者风险感知水平与决策水平关系研究。主要工作和研究内容如下：

运用推演法和假设法，研究了“状态”和事件难度作用下，总体事件相关电位脑成分诱发水平在风险感知水平和决策水平两者之间存在的作用原理和关系。引入总体事件相关电位脑成分诱发水平这一概念，从理论方面提出上述三者之间存在的因素包括：风险感知结果、决策结果、风险感知多个事件相关电位脑成分诱发水平和决策过程中多个事件相关电位脑成分诱发水平存在天秤作用原理和内在关系假设。构建了“风险感知水平-总体事件相关电位脑成分诱发水平-决策水平”的关系概念模型。

应用ERP实验分析方法，研究了基层管理者风险感知与决策过程中诱发出的显著脑成分特征。引入ERP原理与技术，结合主成分分析方法，以“事件难度”为分组，以“状态”为对比条件，识别出量化“风险感知过程中多个事件相关电位脑成分诱发水平”、“决策过程中多个事件相关电位脑成分诱发水平”和“总体事件相关电位脑成分诱发水平的指标”包括：MMN成分的峰值与P300、FRN/ERN、CNV等成分的峰值和潜伏期。进一步发现多外在表征的定量监测问题可转化为量化“多个事件相关电位脑成分诱发水平”的问题进行研究现象。

运用风险感知和多准则决策等理论，提出煤矿基层管理者的“异元异构”多变量的快速、同时量化方法。在筛选出的量化指标基础上，对风险感知水平和决策水平2个变量以及上述3个变量构建了指标体系；结合CMAC神经网络理论和人工鱼群算法，构建了基于CMAC神经网络理论的多变量量化分析模型，发现了“状态”和“事件难度”同时作用下，“状态”对煤矿基层管理者在应对低和高难度事件时的影响弱，“状态”对煤矿基层管理者应对中难度事件时的消极影响强的规律。

运用关联分析、三维非线性拟合、多维比例和非线性回归分析等方法，研究了煤矿基层管理者风险感知与决策水平和总体事件相关电位脑成分诱发水平存在多对一映射特征与其出现疲劳表征的判别规律。发现了基层管理者风险感知水平的比值比总体事件相关电位脑成分诱发水平的比值高时，则其在风险感知和决策过程中，存在疲劳表征或事件难度超出其能力范围的规律，提出了判别疲劳表征的步骤包括：监测风险感知和决策过程中的多个脑成分特征和行为数据、量化多个变量和对比变量的比值等。

上述工作与主要研究内容，进一步拓展了风险感知与决策管理对风险感知与决策水平的定量监测的理论与研究方法，有针对性的提出了基层管理者在风险感知与决策过程中的疲劳表征判别方法，有利于提升基层管理者对风险感知与决策水平的认知层次和管理水平，对煤矿管理失误和决策失误的预防具有重要的理论与实践意义。

The rapid monitoring of fatigue and unfatigued states of mining team-leaders' risk perception level and decision making level is a difficult management-problem in mining. However, based on the psychological scale measurement paradigm, the method of monitoring and identification of the level of risk perception and decision-making level of coal mining team-leaders' has few limitations such as difficulty in data acquisition and delay of measurement results, which has become the bottleneck of monitoring and identifying the fatigue risk perception and decision-making level of mining team-leaders'. Thus, this research takes the mining team-leaders as the research object, taking the cases of different difficulty events as the stimulating conditions, and taking the fatigue and normal states as the comparison group. The fatigue state and normal state of the mining team-leaders have been analysed by adopting method of combining theoretical analysis with event-related potential ERPs experiment. The results achieved are as follows:

A conceptual model of the correlation between risk perception level and ERP event-related potential brain component evoked characteristics and decision level of mining team-leaders has been constructed. Taking the level of risk perception as the independent variable, the level of decision as the dependent variable, and the level of comprehensive induction of the total-components evoked level of event-related potentials as the mediating variable. The results show that: the conceptual model of the corresponding relationship between the risk perception level and the decision making level of mining team-leaders in mining enterprise and the inherent variables has been deduced. 

The risk perception and decision making ERP experimental system for the fatigue and non-fatigue states of the team-leaders in mining enterprise has been designed and tested. The characteristics of MMN components, P300 components, FRN/ERN components, CNV, P300 and FRN/ERN brain components, the risk perception and decision-making level induced by mining team-leaders in mining enterprises were analyzed.The results show that: The level of risk perception is related to the peak value of MMN component, peak value and latency of P300 component, peak value and latency of CNV component, peak value and latency of FRN / ERN component and the difficulty of the event and fatigue status, and can be used as an indicator of the total-component evoked level of ERP.

A multivariate quantitative analysis model based on CMAC neural network theory has been established.The theory of risk perception and muti-criteria of decision-making have been used for combined with the ERP monitoring experiment system. The risk perception level, The level of multiple ERP brain components of risk perception and decision-making and the level of multiple ERP brain components induced during decision making have been selected by five variables. The total-component evoked level of Event-related potential were constructed to a indicator system. The CMAC neural theory and methods, combined with artificial fish school algorithm method are used for establish the muti-variable quantitative analysis model.The risk perception level, multiple ERP components evoked levels of risk perception and decision-making, etc. has been analyst. The results show that: The muti-variable quantitative analysis model based on the CMAC neural network theory can provide a new quantitative method for the rapid quantification of muti-index and muti-variables in the process of coal mine management.

The relationship between the total components evoked level of event-related potential, the risk perception level and decision making level of team-leaders at different difficulty levels in mining enterprise has been revealed.The relationship between risk perception level, decision-making level and the total components evoked level of event-related potential of mining team-leaders have been analyzed. Management theory, combined with 3D nonlinear fitting and the Multi-dimensional scaling method, The relationship between above 3 various have been analyst. The results show that: The relationship between risk perception level and decision-making level, The total components evoked level of event-related potential can be the Carrier to Numerical rendering the risk perception level and the decision-making level have been calculated. The results show that: it is of great guiding significance to guide the efficient decision-making of mining team-leaders.

The above work and main research contents can provide reference for the judgment of whether there is fatigue representation or event difficulty beyond the ability range in the risk perception and decision-making process of mining team-leaders'.