联系我们

论文查询结果	返回搜索

论文编号：	7222
作者编号：	1120120764
上传时间：	2015/6/2 18:01:54
中文题目：	基于情景分析的群体性突发事件演化模型研究
英文题目：	Research on Evolutionary Model of Unexpected Incidents Involving Mass Participation Based on Scenario Analysis
指导老师：	李勇建
中文关键字：	群体性突发事件；随机Petri网；演化博弈；复杂网络；情景分析
英文关键字：	Unexpected incident involving mass participation; Stochastic Petri nets; Evolutionary game; Complex network; Scenario analysis
中文摘要：	当前我国社会正处于转型期，随着改革开放的深入和社会转型的加速，我国社会正步入群体性突发事件的多发阶段。群体性突发事件的发展演化具有高度复杂性，它的频繁发生及其越来越明显的对抗形式引起社会各界的广泛关注，已经成为影响社会稳定的一个重要因素。正确识别和深刻分析群体性突发事件的发展演化机理，对于预防和处理群体性突发事件，减少社会损失、维护社会和谐稳定具有重要意义。基于上述背景，本论文运用管理学、运筹学、社会学、演化博弈理论和复杂网络理论等相关理论和方法，研究了群体性突发事件的发展演化机理问题。具体来讲，本论文研究内容与创新点主要包括以下四部分： （一）群体性突发事件的演化情景分析 本文采用了多案例分析方法提取群体性突发事件的属性，从“事件类型、关键属性、从属属性、环境属性和危害评估属性”对其进行结构化描述。在此基础上，将群体性突发事件演化过程分为潜伏阶段、诱发阶段、发展阶段、高潮阶段、消亡阶段等五个阶段，并分析群体性突发事件演化过程中的相关属性，利用基于模糊集和随机Petri网的建模分析方法，根据随机Petri网与马尔科夫链的同构关系，构建了群体性突发事件演化的随机Petri网模型和等价马尔科夫链模型。以“池州事件”为例，通过马尔科夫链及其相关数学方法对池州群体性事件不同演化状态进行了情景推演模拟，并分析其中的均衡状态及其变动规律，研究表明：群体性突发事件是由于不同社会群体的利益博弈加剧了社会矛盾激化，当社会结构性压力超过社会承受阈值时，在导火索事件诱发下群体行为爆发的过程；同时，群体情绪的相互感染和结构性传导，使群体社会认知产生偏差，导致行为缺乏理性，最终促成群体性突发事件发生。 （二）群体性突发事件异质群体的演化博弈模型 为了分析不同社会群体的利益博弈问题，基于演化博弈理论研究了群体性突发事件中两类异质群体即强势群体与弱势群体策略选择的演化过程，建立了群体性突发事件的演化博弈模型，分析了弱势群体与强势群体行为策略的稳定性。基于数值仿真对模型进行了情景仿真模拟，研究结果表明：在未引入上级政府惩罚机制情景下，当强势群体采取强硬策略的成本与信誉损失成本超过其获得收益及对采取抗争策略的弱势群体惩罚成本之和，并且弱势群体通过抗争获得收益小于其采取抗争成本时，两个异质群体将最终选择合作策略；当强势群体采取强硬策略的收益超过其行动成本、信誉损失与提供补偿成本之和，且弱势群体通过抗争获得收益超过其行动成本、获得补偿及支付惩罚成本时，两个异质群体将选择强硬抗争策略；两群体策略演化速度与策略选择初始比例有直接关系，在初始状态选择策略比例相同情况下，弱势群体均比强势群体更快演化至均衡策略。在引入上级政府惩罚机制情景下，当上级政府惩罚力度高于强势群体采取强硬策略获得的收益与其行动成本、信誉损失成本和对弱势群体补偿成本之差值，且同时高于弱势群体采取抗争策略获得收益与其行动成本差值时，两个异质群体都将放弃强硬对抗策略，最终选择合作策略；随着施加的惩罚增大，对弱势群体策略演化不再显著，而对强势群体策略演化影响却显著增加。 （三）不同社会网络结构下群体性突发事件的演化博弈模型 考虑到不同社会网络结构下群体性突发事件中群体策略选择的演化问题，基于复杂网络和演化博弈理论构建了群体性突发事件社会网络演化博弈模型，分析了不同社会网络的拓扑结构对个体策略选择与行为模式的影响。以弱势群体社会网络为例，研究了社会网络中节点之间的博弈策略演化过程，基于WS小世界网络、BA无标度网络两种复杂网络对弱势群体社会网络上的演化博弈进行了分析。最后，通过数值仿真对网络中个体博弈与群体结构的协同演化模型进行情景仿真模拟，研究结果表明：当网络中个体说服其邻居采取抗争策略获得额外收益大于其说服成本，且其邻居接受说服获得收益大于不接受说服获得收益条件下，随着WS小世界网络重连概率增大，即网络异质性越大，其演化至均衡策略时间越短；在BA无标度网络上，随着网络初始节点数和新节点连接边数的增大，网络聚类系数越大，其演化至均衡策略时间越长；异质网络中节点度大的个体，具有较大影响力，更容易说服带动周围个体接受其策略，易结成联盟采取抗争策略，形成羊群效应，造成更大危害；高度异构的BA无标度网络策略的演化时间明显小于WS小世界网络的演化，BA无标度网络较WS小世界网络更容易引发群体性事件。 （四）不确定环境下群体性突发事件的随机演化博弈模型 针对群体性突发事件在不确定环境下的演化问题，基于演化博弈理论研究了群体性突发事件中强势群体与弱势群体策略选择的演化过程，依据复制动态方程得到了两个群体的行为演化规律。考虑到群体性突发事件演化过程中的随机扰动，引入高斯白噪声来反映群体性突发事件演化过程中受到的随机干扰，建立了不确定环境下群体性突发事件的随机演化博弈模型，分析了弱势群体与强势群体行为策略的稳定性。运用随机Taylor展开理论和It?型随机微分方程对模型进行了求解，并对模型进行不同情景仿真模拟，研究结果表明：在不确定环境下，两类异质性群体受随机因素的干扰影响，当采取抗争策略成本较大时，随着白噪声强度减小，弱势群体会较快妥协，采取合作策略；当采取强硬策略获取额外收益较大时，随着白噪声强度增大，强势群体会更倾向于采取强硬策略。结合上述情景仿真结果提出了相应的对策建议，为群体性突发事件“情景-应对”提供应急决策支持。
英文摘要：	With the thorough development of reform and opening-up and acceleration of social transformation, our country is entering a multiple stage of unexpected incidents involving mass participation. Unexpected incidents involving mass participation are highly complex, and their frequent and more obvious forms of confrontation caused widespread concern in the community. So Mass Emergency has become one important factor that influences social stability. Correctly realizing and analysising the evolution mechanism of Mass Emergencies, it is of great significance to prevent and response mass emergencies, reduce social lost, and ensure social harmony and stability. Based on the above backdrop, by adopting theories and methods of Management Science, Operations Research, Sociology, Evolutionary Game Theory and Complex Network theory, this dissertation studies the evolutionary mechanism of unexpected incidents involving mass participation. More specifically, the research and innovation of the dissertation include four aspects as follows: 1. Scenario evolvement analysis of unexpected incidents involving mass participation The dissertation extracts unexpected incident attributes by the approach of multi-case study to describe the unexpected incident structurally from the event type, the key attributes, the secondary attributes, the environment attributes and the hazard assessment attributes. On the basis of the structural description, we divide the evolution of unexpected incidents involving mass participation into five phases such as latent phase, induction phase, development phase, climax phase, and declining phase. Then this dissertation analyzes the properties of related events in the events chain for the process of unexpected incident, apply the fuzzy sets and stochastic Petri nets to model the unexpected incident, and establish the corresponding Markov chain model based on isomorphic relation between stochastic Petri nets and Markov chain. Then the dissertation takes different evolutionary status of mass incidents for scenario derivation simulation and studies the equilibrium state and fluctuation pattern of the system for evaluating and improving the system with Markov chain and corresponding mathematics method in the case of “Chi’zhou incident”, the main conclusions are the followings: mass incidents are a outbreak process of group behavior in induction of the fuse event with the benefits game of different social groups exacerbated social conflicts, when social structural pressures over social tolerance threshold; and the group emotional cross contamination and structural transfer, leading to groups social cognition bias and irrational behavior, then eventually mass incidents occurred. 2. Evolutionary game model of heterogeneous group in unexpected incidents involving mass participation In order to analyze the benefit game problem of different social groups, we study the strategy selection process of two types of social groups, i.e. the social powerful group and social vulnerable group, in the unexpected incidents involving mass participation based on evolutionary game theory. Then this dissertation establishes the evolutionary game model of mass emergencies and analyzes the behavior strategy stability of the social powerful group and social vulnerable group. This dissertation introduces different evolutionary status of mass incidents for scenario derivation simulations, the main conclusions are as followings: Under the scenario without the higher levels of government’s punishment mechanism, when the sum of the costs to take tough policy and credit losses for the social powerful group exceeds its benefits and the punishment costs of the vulnerable groups to adopt the struggle strategy, and the benefit of the social vulnerable groups by taking struggle strategy is less than the cost to carry out the action, then the two heterogeneous groups will ultimately choose cooperative strategies; when the benefits by taking tough policy of the social powerful group exceed the sum of the cost to carry out the action, credit loss and compensation cost, and the benefits by taking struggle strategy of the social vulnerable group exceeds the sum of the costs to carry out the action, compensation cosst and the penalty costs, then the two heterogeneous groups will select the tough struggle strategy. The strategy evolutionary speeds of these two kinds of groups are directly related with the initial proportion of policies selections, and the speed of the social vulnerable group evolutes to balanced strategy is faster than that of the powerful group in the same circumstances tactics proportion. Under the scenario with the higher levels of government’s punishment mechanism, when the punishment of higher levels of government is greater than the difference in value of the benefits of the social powerful groups by taking tough policy and the costs of their actions, credit losses, compensation costs, and also higher than the difference in value of the benefits of the social vulnerable groups by taking struggle strategy and the costs of their actions, the two heterogeneous groups will abandon tough struggle strategy, then chose cooperation strategy; with the punishment increasing, the impact on evolution of vulnerable groups was no longer significant, but the impact on the evolution of strong group was significantly increased. 3. Evolutionary game model of unexpected incidents involving mass participation in different social network structure Taking into account of the evolutionary problem of strategy selection of unexpected incidents involving mass participation in different social network structure, this dissertation builds a evolutionary game model of the mass emergency in social network based on complex networks and evolutionary game theory, analyzed the influence of individuals strategies selection and behavior pattern affected by the topological structure of different social networks. Then study the evolution of the game strategy between the nodes in the network according to the social network characteristics of vulnerable group, and analyze evolutionary game of vulnerable social networks based on the two complex network-WS small-world networks and BA scale-free networks. Finally, we take the co-evolutionary model of individual strategy and group structure in social network for scenario derivation simulation by numerical simulation, the results show that: Firstly, under the situation of the benefits which individuals of social network get by convincing their neighbors to take struggle strategy is greater than their costs of convincing, and their neighbors get more benefits when they accept the convincement rather than refuse it. and with the increasement of rewiring probability of WS small-world network, that means the heterogeneity of the network will be greater, and the time of equilibrium strategy evolution will be shorter; Secondly, on BA scale-free network, with the increasing number of initial network nodes and the new node connected to sides, the greater that the average degree of the network will be, the longer the equilibrium strategy evolutionary time will be. Thirdly, the larger node degree of individual in heterogeneous network is of great influence, and it is easier to convince an individual to accept its strategy to make an alliance adopting resisting strategy, and then make the formation of herding to cause greater harm. The evolutionary time of highly heterogeneous BA scale-free network is significantly less than the WS small-world networks, so BA scale-free network becomes more easily to cause mass incidents than the WS small-world network. 4. Stochastic evolutionary game model of unexpected incidents involving mass participation under uncertain environment In order to solve the evolutionary problems of unexpected incidents involving mass participation under uncertain environment, the strategy developing process of social powerful group and vulnerable group in the unexpected incidents is considered based on evolutionary game, and the behavior evolutionary rule of two groups is developed by dynamic replication equation. Taking into account the evolution of mass incidents of random disturbance, the Gaussian white noise drew to reflect random disturbance. The dissertation builds a stochastic evolutionary game model of unexpected incidents involving mass participation under uncertain environment, and analyzes the behavioral stability of social powerful group and vulnerable group. The model is solved by the stochastic Taylor expanded theory and It? stochastic differential equations theory, and take different evolutionary status of mass incidents for scenario derivation simulation, the main conclusions were the followings: the social system is effected by random disturbance under uncertain environment, when social vulnerable group take the larger cost of struggle strategy with the intensity of white noise decreased rapidly, which would take cooperation strategy; when powerful group obtain the larger additional revenue of strategy with the intensity of white noise increased rapidly, which would take tough strategy. Finally, this dissertation provides decision-making support for responding to the unexpected incidents involving mass participation in “scenario-response”.
查看全文：	预览  下载(下载需要进行登录)