采煤工作面作为煤矿第一生产现场，是事故的多发地点。为科学预防控制采煤工作面事故的发生，就必须采取相应的管控措施对事故背后的高概率险兆事件进行管理。但目前关于高概率险兆事件的管理研究尚在起步阶段，因此本研究从复杂适应系统的角度出发，构建了采煤工作面高概率险兆事件 Multi-agent system（MAS）仿真模型，应用计算机技术对模型进行了仿真模拟，分析了采煤工作面高概率险兆事件的演化规律，为采煤工作面高概率险兆事件的管理提供理论基础与实践指导。论文的主要研究内容及结论如下：

（1）开展了采煤工作面高概率险兆事件的理论研究。明确了采煤工作面高概率险兆事件的概念，结合三类危险源理论构建了采煤工作面高概率险兆事件致因模型，并对采煤工作面高概率险兆事件演化系统进行复杂适应性特征分析，证明了采煤工作面高概率险兆事件演化系统为复杂适应系统；根据主体划分的基本原则，将采煤工作面高概率险兆事件演化系统主体划分为作业人员 Agent、管理人员 Agent、高层管理Agent、工作面环境 Agent 及设备 Agent；通过文献梳理和专家访谈，从五个主体层面归纳出 22 个主体属性指标，经过问卷调查与结果检验分析筛选出 19 个主体属性，确定了各主体属性间的行为策略及交互规则。

（2）构建了基于 MAS 的采煤工作面高概率险兆事件仿真模型。运用层次分析法计算模型中行为策略与交互规则内相关系数的权重，并根据相关文献、专家访谈对初始模型参数进行了设置；利用 Netlogo 仿真平台编制仿真程序实现 Agent 的建立与可视化，同时针对仿真系统的仿真界面与功能区进行设计，完成了采煤工作面高概率险兆事件MAS 仿真模型的构建；对仿真模型进行了效用检验及灵敏度分析，证明模型具有较好的实用性和可靠性。

（3）开展了采煤工作面高概率险兆事件演化系统仿真模拟研究。通过设定仿真模型的基准模式与受控模式进行对比分析，发现系统内各主体属性对于采煤工作面高概率险兆事件演化风险的影响程度存在差异，其中管理人员的安全教育培训水平和高层管理者的安全重视程度对采煤工作面高概率险兆事件演化风险影响最大；采煤工作面高概率险兆事件的发生是复杂适应系统内各主体属性相互作用而出现的涌现现象，系统内各主体间相互影响，任何主体属性出现异常状态都可能诱发采煤工作面高概率险兆事件，需均衡提升各主体的安全水平，才能有效的避免采煤工作面高概率险兆事件的发生，保障煤矿企业的安全生产。

As the primary production site of coal mines, coal mining face are prone to accidents. In order to scientifically prevent and control the occurrence of coal mining accidents, corresponding control measures must be taken to manage the high probability near-misses behind the accidents. However, at present, the research on the management of high probability near-misses is still in its infancy. Therefore, from the perspective of complex adaptive systems, this study constructed a Multi Agent System (MAS) simulation model for high probability near-misses in coal mining face. Computer technology was applied to simulate the model and analyze the evolution law of high probability near-misses in coal mining face, providing theoretical basis and practical guidance for the management of high probability near-misses in coal mining face. The main research contents and conclusions of the paper are as follows:

(1) The theoretical research on high-probability near-misses in coal mining face was carried out. The concept of high probability near miss event in coal mining face is defined, the causal model of high probability near-misses in coal mining face is constructed by combining the theory of three types of hazard, and the complex adaptive characteristics of the high-probability near-misses evolution system of coal mining face are analyzed, which proves that the high-probability near-misses evolution system of coal mining face is a complex adaptive system. According to the basic principle of agent division, the main agent of the high-probability near-misses evolution system in coal mining face is divided into operator agent, manager agent, senior manager agent, workplace environment agent, and device agent. Through literature combing and expert interviews, 22 agent attribute indicators were summarized from the five agent levels, 19 agent attributes were screened out through questionnaire survey and result test analysis, and the behavior strategies and interaction rules of each agent attribute were determined.

(2) Simulation model of high probability near-misses in coal mining face based on MASis constructed. Calculate the weight coefficient of the correlation coefficient between behavioral strategies and interaction rules in the model using the analytic hierarchy process, and set the initial model parameters based on relevant literature and expert interviews; Using the Netlogo simulation platform, a simulation program is developed to realize the establishment and visualization of the agent, and the simulation interface and functional area of the simulation system are designed to complete the construction of the MAS simulation model for high probability near-misses in coal mining face; The utility test and sensitivity analysis of the simulation model prove that the model has good practicality and reliability.

(3) Simulate and analyze the results of the evolution system of high probability near-misses in coal mining face. By setting the benchmark mode of the simulation model and comparing it with the controlled mode, the results show that there are differences in the degree of influence of the attributes of various agents within the system on the evolution risk of high probability near-misses in coal mining face, among which the level of safety education and training of managers and the level of safety emphasis of senior managers have the greatest impact on the evolution risk of high probability near-misses in coal mining face; The occurrence of high probability near-misses in coal mining face is an emerging phenomenon that occurs due to the interaction of various agents attributes within a complex adaptive system. The interaction of various agents within the system can lead to high probability near-misses in coal mining face if any abnormal state of agent attributes occurs. It is necessary to balance and improve the safety level of each agents in order to effectively avoid the occurrence of high probability near-misses in coal mining face, ensure the safety production of coal mining enterprises.

[1]中华人民共和国国家统计局. http://www.stats.gov.cn/.

[2]赵百余, 张学亮. 保德煤矿智能化综放工作面安全管控模式及对策[J]. 煤炭科学技术, 2022, 50(S1): 244-249.

[3]谭章禄, 吴琦, 肖懿轩, 等. 基于图元体系的煤炭企业可视化管理场景构建[J]. 科技管理研究, 2019, 39(24): 179-187.

[4]国家矿山安全监察局. https://www.chinamine-safety.gov.cn/.

[5]Heinrich, H. W. Industrial Accident Prevention[M]. McGraw-Hill, 1979.

[6]田水承, 景国勋. 安全管理学[M]. 北京: 机械工业出版社, 2022.

[7]孙庆兰, 田水承, 王艳. 煤矿运输险兆事件综合防控管理研究[J]. 西安科技大学学报, 2019, 39(03): 411-418.

[8]樊梦琰. 煤矿高概率险兆事件犹豫模糊多属性决策分级研究[D]. 西安: 西安科技大学, 2020.

[9]田水承, 寇猛, 金梦. 煤矿水害险兆事件管理评价指标体系构建及其应用[J]. 西安科技大学学报, 2016, 36(02): 181-186.

[10]盛昭瀚, 张维. 管理科学研究中的计算实验方法[J]. 管理科学学报, 2011, 14(05): 1-10.

[11]Skiba. An accident model[J]. Occupational Safety and health 1974, (4): 14-16.

[12]Lisa M. Morrison. Best practices in incident investigation in the chemical process industries with examples from the industry sector and specifically from Nova Chemicals [J]. Journal of Hazardous Materials 111 (2004): 161-166.

[13]Samer A M Nashef. The lancet[R]. 2003. 1(11): 361.

[14]Fabricio Borges Cambraia, Tarcisio Abreu Saurin, Carlos Torres Formoso. Identification, analysis and dissemination of information on near misses: A case study in the construction industry[J]. Safety Science, 2010, 48(1): 91-99.

[15]Aileen R. Killen, Suzanne C. Beyea. Learning from near misses in an effort to promote patient safety[J]. ARON Journal, 2003, 77(2): 423-424.

[16]David B, Siobhan Havis, J. J. Payne-James. Near miss incidents in police custody suites in London in 2003: A feasibility study[J]. Journal of Clinical Forensic Medicine, 2006, 13: 60-64.

[17]Marsh P, Kendrick D. Near miss and minor injury information-can it be used to plan and evaluate injury prevention programmes[J]. Accident Analysis & Prevention, 2000, 3: 345-354.

[18]周志鹏, 李启明, 邓小鹏, 等. 险兆事件管理系统在地铁施工安全管理中的应用[J]. 解放军理工大学学报(自然科学版), 2009, 10(06): 597-603.

[19]邓小鹏, 周志鹏, 李启明, 等. 地铁工程Near-miss知识库构建[J]. 东南大学学报(自然科学版), 2010, 40(05): 1103-1109.

[20]田卫, 李慧民, 闫瑞琦, 等. 基于Near-Miss的高速公路专项养护工程安全管理模式[J]. 西安建筑科技大学学报(自然科学版), 2013, 45(04): 548-553+558.

[21]杨禄. 煤矿险兆事件管理信息系统研究[D]. 西安: 西安科技大学, 2014.

[22]田水承, 孙曙英, 于观华, 等. 煤矿险兆事件主动上报意愿影响因素研究[J]. 西安科技大学学报, 2014, 35(5): 517-522.

[23]Linda Wright, Van der Schaaf T. W. Accident versus near miss causation:a critical review of the literature, an empirical test in the UK railway domain,and their implications for other sectors[J]. Journal of Hazardous Materials, 2004, 111(111): 105-110.

[24]Marieke Kessels-Habraken, Tjerk Van der Schaaf, Jan De Jonge, et al. Defining near misses: Towards a sharpened definition based on empirical data about error handling processes[J]. Social Science&Medicine, 2010, 70: 1301-1308.

[25]张晓. Near-Miss管理分析及其在某公司的应用研究[D]. 北京: 中国地质大学, 2007.

[26]崔文倩. 我国企业险兆事件管理方案设计[D]. 北京: 中国地质大学, 2011.

[27]于观华. 基于三类危险源的煤矿险兆事件管理研究[D]. 西安: 西安科技大学, 2013.

[28]高瑞霞. 煤矿火灾险兆事件作用机制的调节效应分析[J]. 煤矿安全, 2016, (03): 218-221.

[29]孙庆兰, 田水承. 基于扎根理论的煤矿运输险兆事件影响因素研究[J]. 安全与环境学报, 2017, 17(03): 1031-1037.

[30]马晓雪, 刘雨, 何佩龙, 等. 船舶险兆事件主动报告影响因素[J]. 中国航海, 2021, 44(01): 8-13.

[31]于观华, 田水承, 王莉, 等. 基于三类危险源的掘进工作面瓦斯爆炸模糊综合安全评价[J]. 矿业安全与环保, 2013, 40(1): 111-114.

[32]陈霞, 戴广龙. 煤矿未遂事件(Near-Miss)管理研究[J]. 煤矿安全, 2014, 45(7): 236-237.

[33]金梦. 煤矿水害险兆事件的影响因素和管理措施[D]. 西安: 西安科技大学, 2015.

[34]田水承, 樊梦琰, 韩旭, 等. 煤矿险兆事件风险评价指标及分级方法[J]. 矿业安全与环保, 2021, 48(04): 123-127+134.

[35]Eldar, Nadyrov, Alexander Rozhko, Viacheslav Kravtsov, et al. Karyopathological traits of thyrocytes and exposure to radioiodines in Belarusian children and adolescents following the accident at the Chernobyl nuclear power plant[J]. Radiation and environmental biophysics, 2012, 51(02): 93-187.

[36]Eva-Maria Merz, Ezgi Ozeke-Kocabas, Frans J Oort, et al. Intergenerational family solidarity: Differences between immigrant groups and generations[J]. Journal of Family Psychology, 2009, 23(03): 291-300.

[37]SUZUKI T, AOKI Y, KATAOKA H. Pedestrian near-miss analysis on vehicle-mounted driving recorders [C]//Fifteenth Iapr International Conference on Machine Vision Applications, 2017, 416-419．

[38]张涛伟, 田水承, 李树刚. 基于险兆事件的煤与瓦斯突出灾害概率模型构建[J]. 煤矿安全, 2014, 45(03): 156-159.

[39]田水承, 梁清, 马文赛, 等. 煤矿瓦斯爆炸险兆事件致因模型构建[J]. 煤矿安全, 2017, 48(04): 226-229.

[40]田水承, 杨鹏飞, 唐凯, 等. 煤矿高概率险兆事件界定及其研究意义[J]. 西安科技大学学报, 2020, 40(04): 566-571.

[41]李润求, 施式亮, 彭新. 煤矿瓦斯爆炸事故演化的突变模型[J]. 中国安全科学学报, 2008(03): 22-27.

[42]龚承柱, 李兰兰, 柯晓玲, 等. 基于multi-agent的煤矿水害演化模型[J]. 煤炭学报, 2012, 37(06): 1005-1009.

[43]张津嘉, 许开立, 王贝贝, 等. 瓦斯爆炸事故演化机理的综合论事故模型研究[J]. 中国安全科学学报, 2015, 25(04): 53-57.

[44]张津嘉, 许开立, 王贝贝, 等. 瓦斯爆炸事故风险耦合演化机理研究[J]. 中国安全科学学报, 2016, 26(03): 81-85.

[45]孟祥坤, 陈国明, 朱红卫. 海底管道泄漏风险演化复杂网络分析[J]. 中国安全生产科学技术, 2017, 13(04): 26-31.

[46]陈国华, 王永兴. 基于系统动力学的化工园区事故演化仿真[J]. 科学技术与工程, 2018, 18(19): 347-352.

[47]庄文英, 许英姿, 任俊玲, 等. 突发事件舆情演化与治理研究——基于拓展多意见竞争演化模型[J]. 情报杂志, 2021, 40(12): 127-134+185.

[48]李小莉, 曹策俊, 张帆顺. 进口国规制下医药产品出口安全监管演化仿真[J]. 系统仿真学报, 2021, 33(09): 2252-2260.

[49]杜元伟, 孙浩然, 王一凡, 等. 海洋牧场生态安全监管的演化博弈模型及仿真[J]. 生态学报, 2021, 41(12): 4795-4805.

[50]孙彧, 曹雷, 陈希亮, 等. 多智能体深度强化学习研究综述[J]. 计算机工程与应用, 2020, 56(05): 13-24.

[51]Franco Cicirelli, Andrea Giordano, Libero Nigro. Efficient environment management for distributed simulation of large-scale situated multi-agent systems[J]. Concurrency & Computation Practice & Experience, 2015, 27(03): 610-632.

[52]Gilbert N, Terna P. How to build and use agent-based models in social science[J]. Mind & Society: Cognitive Studies in Economics and Social Sciences, 2000, 1(1): 57-72.

[53]Stefan Woltmann, Maximilian Zarte, Julia Kittel, et al. Agent Based Simulation Model of Virtual Power Plants for Greener Manufacturing[J]. Procedia CIRP, 2018, 69: 377-382.

[54]Shobole Abdulfetah Abdela, Wadi Mohammed. Multi-agent systems application for the smart grid protection[J]. Renewable and Sustainable Energy Reviews, 2021, 149: 111352

[55]蔡林沁, 罗志勇, 王颋, 等. 基于Multi-agent的煤矿虚拟环境体系建模[J]. 煤炭学报, 2010, 35(01): 61-65.

[56]李乃文, 黄俊婷, 牛莉霞. 基于Multi-agent的矿工违章行为演化模型[J]. 中国安全科学学报, 2013, 23(11): 10-15.

[57]聂兴信, 白存瑞, 张婧静. 基于ABMS的矿山监管Agent工作倦怠现象的仿真研究[J]. 中国安全生产科学技术, 2019, 15(12): 156-162.

[58]陈培珠, 陈国华, 周利兴, 等. 化工园区多Agent协同应急智能决策体系[J]. 化工进展, 2021, 40(08): 4656-4665.

[59]王康勃, 侯岳, 龚立, 等. 基于多智能体的舰艇消防虚拟仿真训练系统[J]. 船舶工程, 2021, 43(04): 55-59+76.

[60]潘理虎, 杨芬玉, 芦飞平, 等. 街道视角的多智能体城市安全宜居度建模[J]. 应用科学学报, 2022, 40(01): 47-60.

[61]Cheng Lianhua, Guo Huimin, Lin Haifei. Evolutionary model of coal mine safety system based on multi-agent modeling[J]. Process Safety and Environmental Protection, 2021, 147: 1193-1200.

[62]倪建军, 徐立中, 王建颖. 基于CAS理论的多Agent建模仿真方法研究进展[J]. 计算机工程与科学, 2006, 05: 83-86+97.

[63]霍兰, 周晓枚, 韩晖. 隐秩序: 适应性造就复杂性[M]. 上海: 上海科技教育出版社, 2000.

[64]Taylor-Swanson Lisa, P. Tanuja, C. Lisa. Complex Adaptive Systems Theory and Inter-Rater Reliability: Proposed Answers to Challenging Questions[J]. Journal of alternative and complementary medicine, 2019, 25(11): 1074-1076.

[65]陈禹. 复杂适应系统(CAS)理论及其应用——由来、内容与启示[J]. 系统辩证学学报, 2001(04): 35-39.

[66]葛永林, 徐正春. 论霍兰的CAS理论——复杂系统研究新视野[J]. 系统辩证学学报, 2002(03): 65-67+75.

[67]范如国. 平台技术赋能、公共博弈与复杂适应性治理[J]. 中国社会科学, 2021(12): 131-152+202.

[68]赖冉. 基于CAS理论的建筑工人不安全行为研究[D]. 大连: 大连理工大学, 2021.

[69]王雪萍, 高新勤. 复杂层次模型中知识表达与推理方法研究[J]. 系统工程理论与实践, 2019, 39(11): 2918-2927.

[70]赵运平. 基于复杂适应系统交叉融合解蔽新工科建设堆题方略[J]. 系统科学学报, 2023, 31(01): 44-49.

[71]范玉顺, 曹军威. 多代理系统理论、方法与应用[M]. 北京: 清华大学出版社, 2002: 1-4.

[72]Wooldridge M, Jennings N R. Intelligence agents: Theory and practice[J]. Knowledge Engineering Review, 1994, 10(2): 115-152.

[73]杜威, 丁世飞. 多智能体强化学习综述[J]. 计算机科学, 2019, 46(08): 1-8.

[74]孙尊涛. 基于多Agent的企业生态群落仿真研究与实现[D]. 上海: 复旦大学, 2011.

[75]田水承, 周可柔, 杨雪健, 等. 险兆事件视域下个体和组织事故3道“防火墙”[J]. 中国安全科学学报, 2018, 28(05): 62-67.

[76]汪佑武. 采煤概论[M]. 北京: 煤炭工业出版社, 2008.

[77]田水承, 李红霞, 王莉. 3类危险源与煤矿事故防治[J]. 煤炭学报, 2006(06): 706-710.

[78]田水承, 黄权. 基于文本挖掘的煤矿瓦斯事故致因分析[J]. 煤矿安全, 2022, 53(05): 241-245.

[79]成连华. 煤矿企业安全系统进化机制研究[D]. 西安: 西安科技大学, 2015.

[80]Health and Safety Execute(HSE). Reducing error and influencing behaviour[M]. HSG48. UK. HSE Books, 1999.

[81]牛莉霞, 刘谋兴, 李乃文, 等. 基于ABMS的矿工习惯性违章行为演化仿真[J]. 中国安全科学学报, 2015, 25(11): 34-40.

[82]陈悦峰, 董原生, 邓立群. 基于Agent仿真平台的比较研究[J]. 系统仿真学报, 2011, 23: 110-116.

[83]韩心星, 李舒. 基于Netlogo的大型客运汽车人员疏散模拟[J]. 科学技术与工程, 2022, 22(22): 9890-9895.