【摘要】：近年來,鐵路突發(fā)事件時有發(fā)生,對鐵路運輸生產(chǎn)和經(jīng)濟運行造成了巨大損失。隨著鐵路系統(tǒng)新技術(shù)及新裝備的應(yīng)用,系統(tǒng)之間的關(guān)聯(lián)度、耦合性和復(fù)雜性也不斷提高,加劇了鐵路突發(fā)事件應(yīng)對的緊迫性和復(fù)雜性�；陬A(yù)案管理的傳統(tǒng)應(yīng)急管理模式在鐵路突發(fā)事件的應(yīng)急中遇到了前所未有的挑戰(zhàn)。鐵路應(yīng)急管理正在由預(yù)案管理向情景管理轉(zhuǎn)變。深入研究基于情景應(yīng)對的鐵路應(yīng)急決策模型,構(gòu)建鐵路突發(fā)事件的情景描述體系和情景重構(gòu)模型理論方法具有重要的現(xiàn)實意義和應(yīng)用價值。本文在總結(jié)對國內(nèi)外相關(guān)研究成果的基礎(chǔ)上,以鐵路突發(fā)事件情景的組成結(jié)構(gòu)為出發(fā)點,綜合運用本體論、形式化理論及定性與定量相結(jié)合的方法,建立了面向鐵路應(yīng)急決策的突發(fā)事件情景構(gòu)建框架,深入研究了情景模型的構(gòu)建方法、情景信息的集成方法、應(yīng)急決策方法以及情景推演等關(guān)鍵問題。首先,本文在鐵路應(yīng)急管理體系下,對突發(fā)事件應(yīng)急管理的特點和流程進行了深入分析,以情景分析理論為基礎(chǔ)對情景要素進行分類。通過對情景要素的本體設(shè)計,建立鐵路突發(fā)事件情景分層網(wǎng)絡(luò)模型。引入情景重構(gòu)的思想,研究鐵路應(yīng)急情景重構(gòu)的流程及方法,并對情景網(wǎng)絡(luò)中節(jié)點的依賴關(guān)系和可視化的表達進行了描述和分析。其次,分析情景構(gòu)建的信息需求以及信息的來源和特征,針對鐵路應(yīng)急環(huán)境對信息服務(wù)集成的自治性的需求,研究了基于SOAACM的鐵路應(yīng)急情景信息服務(wù)集成方法。該方法通過引入Agent技術(shù),采用軟件Agent代理Web服務(wù),將全局定義的情景模型劃分為各個部門的本地流程模型,使得信息集成服務(wù)的控制分布到多個服務(wù)結(jié)點。與基于BPEL4WS的集成模型相比,使用該方法構(gòu)造的信息服務(wù)集成系統(tǒng)具有更好的自治性和可伸縮性。再次,研究基于情景網(wǎng)絡(luò)分析的鐵路應(yīng)急決策方法,將情景模型與粗糙集理論、D-S證據(jù)理論和貝葉斯網(wǎng)絡(luò)有機結(jié)合完成鐵路應(yīng)急決策的推理。本文從實際應(yīng)用的角度出發(fā),依托情景要素之間的關(guān)聯(lián)關(guān)系,融合專家知識構(gòu)建貝葉斯網(wǎng)絡(luò)模型。為了降低情景網(wǎng)絡(luò)輸入?yún)?shù)的復(fù)雜性,采用基于信息熵的粗糙集知識約簡方法實現(xiàn)了最小決策信息集的提取。實例分析結(jié)果表明情景網(wǎng)絡(luò)借助于簡化后的模型進行決策推理,算法實現(xiàn)方便,即使在信息不完備的情況下也能得到滿意的結(jié)果,更適合于實際應(yīng)用。然后,建立了基于事件驅(qū)動的鐵路突發(fā)事件情景推演模型。依據(jù)鐵路突發(fā)事件情景演變的過程,將突發(fā)事件情景劃分為初始情景、中間情景和結(jié)束情景,并對其相互關(guān)系進行分析。結(jié)合貝葉斯網(wǎng)絡(luò)推理技術(shù),詳細描述了鐵路突發(fā)事件情景推演過程。以列車脫軌事故為例,對情景網(wǎng)絡(luò)的節(jié)點變量狀態(tài)進行推演,推演結(jié)果與現(xiàn)實情況基本一致,證明該方法的可行性和有效性。最后,本文對鐵路應(yīng)急情景重構(gòu)與輔助決策原型系統(tǒng)RESRADPS進行了設(shè)計,詳細設(shè)計了系統(tǒng)的物理和邏輯架構(gòu)以及功能模塊結(jié)構(gòu),基于具體的軟硬件平臺實現(xiàn)了原型系統(tǒng),并結(jié)合對系統(tǒng)功能界面和運行結(jié)果的分析,驗證了本文所提出的理論、方法以及所構(gòu)建系統(tǒng)的正確性、可行性和有效性。
[Abstract]:Railway emergencies have occurred frequently in recent years, which have caused great losses to railway transportation production and economic operation. With the application of new technologies and equipment in railway system, the correlation, coupling and complexity between the systems are also increasing, which aggravates the urgency and complexity of dealing with railway emergencies. Emergency management mode is facing unprecedented challenges in railway emergency management. Railway emergency management is changing from plan management to scenario management. On the basis of summarizing the relevant research results at home and abroad, this paper, starting from the composition structure of railway emergency scenarios, integrates ontology, formalization theory and qualitative and quantitative methods, establishes the framework of emergency scenarios for railway emergency decision-making, and studies the scenarios in depth. First of all, this paper analyzes the characteristics and process of emergency management under the railway emergency management system, and classifies the scene elements based on the theory of scenario analysis. A hierarchical network model of railway emergency scenarios is established by using the concept of scenario reconstruction. The process and method of railway emergency scenario reconstruction are studied, and the dependency and visualization of nodes in the network are described and analyzed. In order to satisfy the autonomous requirement of railway emergency environment for information service integration, the integration method of railway emergency scenario information service based on SOAACM is studied. By introducing agent technology and adopting software agent proxy Web service, the scenario model defined globally is divided into local process models of each department, which makes the information integration service possible. Compared with the integrated model based on BPEL4WS, the information service integration system constructed by this method has better autonomy and scalability. Thirdly, the railway emergency decision-making method based on scenario network analysis is studied, which combines scenario model with rough set theory, D-S evidence theory and Bayesian network. In order to reduce the complexity of input parameters of scenario network, a knowledge reduction method based on information entropy is used to realize the minimum decision information set extraction. The results of case analysis show that the scenario network can make decision reasoning by means of the simplified model, and the algorithm is easy to implement. Even if the information is incomplete, the algorithm can get satisfactory results, which is more suitable for practical application. Then, the scenario deduction model of railway emergency based on event-driven is established. The evolvement process divides the emergency scenarios into initial scenario, intermediate scenario and end scenario, and analyzes their relationship. Combined with Bayesian network reasoning technology, the process of railway emergency scenario deduction is described in detail. Finally, this paper designs RESRADPS, which is a prototype system of railway emergency scene reconfiguration and auxiliary decision-making. It designs the physical and logical framework and function module structure of the system in detail. The prototype system is implemented based on the specific hardware and software platform, and combined with the system power. The correctness, feasibility and validity of the proposed theory, method and the system are verified by the analysis of energy interface and operation results.
【學(xué)位授予單位】：蘭州交通大學(xué)
【學(xué)位級別】：博士
【學(xué)位授予年份】：2014
【分類號】：U298

【共引文獻】

相關(guān)碩士學(xué)位論文 前1條

1 冒朝靜;山區(qū)高速公路分離式路基段突發(fā)事件應(yīng)急管理資源配置研究[D];重慶交通大學(xué);2013年

，

本文編號：2226629