矿工安全行为的群体演化和干预策略研究:基于多智能体仿真分析

Study on the Group Evolution and Intervention Strategies of Miners′Safety Behavior:Based on Multi-agent Simulation Analysis

随着我国多数矿山企业正实现从“系统智能化”向“智能系统化”的跨越,矿工群体逐渐呈现知识化、年轻化的特征。为推进矿工安全行为管理,结合后悔理论构建了包含情绪反应的矿工安全行为决策模型,并制定了考虑矿工风险偏好异质性的互动规则,通过多智能体仿真的方法,探讨了正念和警示两类事前干预策略下矿工群体安全行为的演化路径。结果表明,仅靠处罚这一事后控制手段并不能有效促进矿工选择安全行为;在事前干预策略中,正念策略比警示策略更有利于促进矿工的安全行为,且这两种策略具有协同效应。此外,矿工之间的互动强度会影响安全行为决策,连结的邻居数量越多,邻居更新越快,矿工会越倾向选择不安全行为。最后,在陕西榆林某矿山企业开展了现场实验,实证检验了正念策略和警示策略对矿工安全行为的激励作用。

As most of China′s mining enterprises are realizing the leap from"system intelligence"to"intelligent systematization",the miner group gradually presents the characteristics of intellectualization and youthfulness.In order to promote the management of miners′safety behavior,a decision-making model of miners′safety behavior including emotional response is constructed based on regret theory,and the interaction rules considering the heterogeneity of miners′risk preference are formulated.Through multi-agent simulation,the evolution path of miners′group safety behavior under two kinds of pre-intervention strategies of mindfulness and warning is discussed.The results show that punishment alone cannot effectively promote miners′choice of safety behavior.In the pre-intervention strategy,the mindfulness strategy is more conducive to promoting the safety behavior of miners than the warning strategy,and these two strategies have a synergistic effect.In addition,the intensity of interaction between miners will affect the decision-making of safety behavior.The more the number of neighbors connected,the faster the neighbors update,and the more likely the miners will choose unsafe behavior.Finally,a field experiment is carried out in a mining enterprise in Yulin,Shaanxi Province,and the incentive effect of mindfulness strategy and warning strategy on miners′safety behavior is empirically tested.