基于證據理論的安全防范系統評估與優(yōu)化設計
發(fā)布時間：2018-12-16 10:29
【摘要】：安全防范系統在我們的生活中已經無處不在,但是由于缺乏有效的安全防范系統風險評估方法及優(yōu)化設計手段,目前國內大多數安全防范系統只是對安防產品的簡單組合和主觀布局,容易出現過度保護或者重點區(qū)域保護不足等問題。建立一套完整、量化的安全防范系統評估和優(yōu)化設計方法是我國安防領域中的一個核心問題。本文將證據理論應用到安全防范系統的風險評估和優(yōu)化設計中,結合示例模型說明了方法的具體流程,主要創(chuàng)新性研究成果如下:(1)首次將證據理論系統地應用到安全防范系統的風險評估過程中,提出了一套基于信息融合思想的較為完整的評估方法。對安全防范系統進行了二維平面建模,能夠更加有效地刻畫系統結構信息;對平面點的多部件風險信息進行融合,得到最終風險信息,可有效表征該點所受到的綜合防護;且評估結果能夠實現直觀的三維可視化。(2)首次使用證據函數(BPA)來表達平面點的風險信息,能夠更好地處理“不確定”和“不知道”的情況,減少了信息損失。提出了三種BPA生成模型,基于正態(tài)分布的生成模型、基于高斯過程回歸的無參生成模型和基于隸屬度的生成模型。特別是第三種模型為系統人員提供了一種自定義威脅度生成策略的途徑,再將威脅度結合隸屬度函數生成BPA。(3)文中提出了兩種將BPA轉化為概率的方法,用于實現風險評估的概率式決策,包括基于證據關聯系數的概率轉換方法和基于沖突系數的概率轉換方法,通過算例將它們與經典的賭博概率轉換方法進行比較并給出相應的使用建議。(4)針對安全防范系統的最優(yōu)化設計問題,提出了最優(yōu)化問題的構建方法和求解方案,將DIRECT算法和遺傳算法結合起來,優(yōu)勢互補。能夠求解固定資源條件下系統部件的最佳布局,最大程度發(fā)揮系統的防范效能,有利于提高系統的成本-效益比。
[Abstract]:Security prevention systems are everywhere in our lives, but due to the lack of effective risk assessment methods and optimization design methods, At present, most of the domestic security prevention systems are only a simple combination and subjective layout of security products, which are prone to over-protection or insufficient protection in key areas. It is a core problem in the field of security in China to establish a set of complete and quantitative methods for evaluating and optimizing the security system. In this paper, the evidence theory is applied to the risk assessment and optimization design of the security prevention system, and the concrete flow of the method is illustrated with an example model. The main innovative research results are as follows: (1) for the first time, the evidence theory is systematically applied to the risk assessment process of the security prevention system, and a set of relatively complete evaluation methods based on the idea of information fusion are proposed. The two-dimensional plane modeling of the security prevention system can describe the system structure information more effectively, and the multi-component risk information of the plane point can be fused to obtain the final risk information, which can effectively represent the comprehensive protection of the point. The evaluation results can be visualized intuitively. (2) for the first time, the evidence function (BPA) is used to express the risk information of planar points, which can better deal with the situation of "uncertainty" and "not knowing" and reduce the loss of information. Three kinds of BPA generation models are proposed, which are based on normal distribution, non-parametric generation model based on Gao Si process regression and generating model based on membership degree. In particular, the third model provides a way for system personnel to create a self-defined threat degree strategy. Then, two methods of transforming BPA into probability are proposed in this paper, which combines threat degree with membership function to generate BPA. (3). Probabilistic decision-making for risk assessment, including probabilistic conversion methods based on evidence correlation coefficients and conflict coefficients, By comparing them with the classical gambling probability conversion method, the corresponding suggestions are given. (4) for the optimization design problem of the security prevention system, the construction method and the solution scheme of the optimization problem are proposed. DIRECT algorithm and genetic algorithm are combined to complement each other. It can solve the optimal layout of system components under the condition of fixed resources, maximize the effectiveness of the system, and improve the cost-benefit ratio of the system.
【學位授予單位】：上海交通大學
【學位級別】：博士
【學位授予年份】：2015
【分類號】：TP202;D918.2
，
本文編號：2382190
[Abstract]:Security prevention systems are everywhere in our lives, but due to the lack of effective risk assessment methods and optimization design methods, At present, most of the domestic security prevention systems are only a simple combination and subjective layout of security products, which are prone to over-protection or insufficient protection in key areas. It is a core problem in the field of security in China to establish a set of complete and quantitative methods for evaluating and optimizing the security system. In this paper, the evidence theory is applied to the risk assessment and optimization design of the security prevention system, and the concrete flow of the method is illustrated with an example model. The main innovative research results are as follows: (1) for the first time, the evidence theory is systematically applied to the risk assessment process of the security prevention system, and a set of relatively complete evaluation methods based on the idea of information fusion are proposed. The two-dimensional plane modeling of the security prevention system can describe the system structure information more effectively, and the multi-component risk information of the plane point can be fused to obtain the final risk information, which can effectively represent the comprehensive protection of the point. The evaluation results can be visualized intuitively. (2) for the first time, the evidence function (BPA) is used to express the risk information of planar points, which can better deal with the situation of "uncertainty" and "not knowing" and reduce the loss of information. Three kinds of BPA generation models are proposed, which are based on normal distribution, non-parametric generation model based on Gao Si process regression and generating model based on membership degree. In particular, the third model provides a way for system personnel to create a self-defined threat degree strategy. Then, two methods of transforming BPA into probability are proposed in this paper, which combines threat degree with membership function to generate BPA. (3). Probabilistic decision-making for risk assessment, including probabilistic conversion methods based on evidence correlation coefficients and conflict coefficients, By comparing them with the classical gambling probability conversion method, the corresponding suggestions are given. (4) for the optimization design problem of the security prevention system, the construction method and the solution scheme of the optimization problem are proposed. DIRECT algorithm and genetic algorithm are combined to complement each other. It can solve the optimal layout of system components under the condition of fixed resources, maximize the effectiveness of the system, and improve the cost-benefit ratio of the system.
【學位授予單位】：上海交通大學
【學位級別】：博士
【學位授予年份】：2015
【分類號】：TP202;D918.2
，
本文編號：2382190
本文鏈接：http://www.sikaile.net/falvlunwen/fanzuizhian/2382190.html
