【作者】 姜静逸；

【导师】 韩松臣；

【作者基本信息】 南京航空航天大学 ， 交通运输规划与管理， 2010， 硕士

【摘要】 在综合交通运输体系中,民航运输业受天气影响严重,加上事故发生后人员极低的生还率和较大的社会影响,民航安全问题尤为引人注目。机场是处置民航突发事件、实施应急救援工作的主体。因此,借助先进的信息化技术建立稳定高效的机场应急救援系统是提高机场应急救援能力的迫切需要。本文分别对机场应急救援系统中两个关键性问题进行了深入研究,分别为机场应急救援的规模决策问题和救援资源的动态调配问题。机场应急救援的规模决策对于抢救生命财产至关重要。针对目前救援规模决策缺乏合理规则指导的问题,提出了一种基于满意度控制的数据挖掘方法。该方法通过分析机场应急救援的总目标,根据满意度理论构造异样样本识别的满意度评价模型,并将满意度评价模型中的选择函数和拒绝函数植入数据挖掘的Apriori关联规则算法中,在保持样本完好性的同时识别异样数据,控制挖掘进程,最终能够建立有效的机场应急救援规模决策规则库,用以辅助决策者做出更为科学合理的决策。通过算法实例证明,基于满意控制的数据挖掘算法能更加有效地挖掘出合理的应急救援规模决策规则,提高挖掘的准确度和效率。机场应急资源调配系统结构复杂,影响因素多,开放性强,随机及不确定因素多,因此引入Agent技术。鉴于Agent对不断变化的环境的适应性和求解问题的自主性,以及Agent间的相互合作和协调的特性,本文建立了基于多Agent技术的机场应急资源调配系统。首先根据从功能实体到Agent的映射方法,抽象出构成机场应急资源调配多Agent系统的个体Agent;对个体Agent的模型结构、功能模块以及系统工作方式作了深入的分析;在分析个体Agent模型基础上,进一步研究资源调配过程中个体Agent之间的协同工作过程,最终构建基于多Agent系统的机场应急资源调配体系框架。更多还原

【Abstract】 In the comprehensive transportation system, the safety of aviation was usually paid most attention to, because the aviation transportation industry was affected by the weather, which has very little survivorship rate and great influence. The airdrome is the major place where aviation emergency accidents happen and emergency rescue work is carried out. Therefore, it is urgent required of establishing airport emergency rescue system to enhance the ability of airport emergency rescue by advanced information technology. This article does research on two essential issues: the scale decision-making of the airport emergency rescue and the rescue resource allocation.The scale decision-making of the airport emergency rescue operation is extremely important to save the people and property. In allusion to the absence of reasonable rule guidance in the current rescue scale decision-making, this article puts forward a data mining method based on the satisfaction control. According to the overall objective of the airport emergency rescue and the satisfaction theory, the method constructs the satisfaction evaluation model for identifying abnormal sample, and then plants the selection function and the rejection function of the satisfaction evaluation model into Apriori Association Rule Algorithm of the data mining. It tries to identify the abnormal data with keeping the integrity of the sample and control the mining process. Finally, it establishes an effective assistant decision-making rule base of the airport emergency rescue for the decision maker to make a more reasonable decision. The algorithm example shows that it can obtain more reasonable rules of emergency rescue scale decision-making, increasing the mining accuracy and efficiency.Airport emergency resource allocation system has complex structure, many factors, openness, random and strong uncertainty. In view of the Agent characteristics of adaptability to constantly changing environment, autonomy of solving problems, mutual cooperation and coordination between the Agents, this article establishes the rescue resource recombined Multi-Agent System. Firstly, According to a mapping method from the functional entity to the Agent, it abstracts out the allocation of resources constitute the Individual Agent of the Multi-Agent System; analyzes the model structure, functional module and system working model of the Individual Agent; finally, establish the Multi-Agent System framework of the airport emergency rescue resource allocation, Based on the further research of the collaborative work process between the Individual Agents.更多还原