价值工程
價值工程
개치공정
VALUE ENGINEERING
2013年
30期
80-81
,共2页
港口工程%概率树%贝叶斯算法%风险识别
港口工程%概率樹%貝葉斯算法%風險識彆
항구공정%개솔수%패협사산법%풍험식별
port engineering%probability tree%Bayesian algorithm%risk identification
本文的目的在于对港口工程建设风险进行风险识别。由于港口工程涉及风险因素面较广,故在施工过程中进行风险管理有着十分重要的意义,因此本文利用港口工程风险评价指标体系,建立了故障树树模型。由于传统的故障树定量分析中采用的是风险因素本身的先验概率进行重要度分析,对风险因素本身不够准确和合理,故本文引入贝叶斯算法,通过相关专家及施工技术人员提供的现场经验,得出风险因素的重要度权重概率和相关概率,计算出风险因素的后验概率。并在传统的故障树重要度分析中,将贝叶斯概率引入,推导出贝叶斯关键重要度,进而识别出港口工程建设风险中的重要风险因素。
本文的目的在于對港口工程建設風險進行風險識彆。由于港口工程涉及風險因素麵較廣,故在施工過程中進行風險管理有著十分重要的意義,因此本文利用港口工程風險評價指標體繫,建立瞭故障樹樹模型。由于傳統的故障樹定量分析中採用的是風險因素本身的先驗概率進行重要度分析,對風險因素本身不夠準確和閤理,故本文引入貝葉斯算法,通過相關專傢及施工技術人員提供的現場經驗,得齣風險因素的重要度權重概率和相關概率,計算齣風險因素的後驗概率。併在傳統的故障樹重要度分析中,將貝葉斯概率引入,推導齣貝葉斯關鍵重要度,進而識彆齣港口工程建設風險中的重要風險因素。
본문적목적재우대항구공정건설풍험진행풍험식별。유우항구공정섭급풍험인소면교엄,고재시공과정중진행풍험관리유착십분중요적의의,인차본문이용항구공정풍험평개지표체계,건립료고장수수모형。유우전통적고장수정량분석중채용적시풍험인소본신적선험개솔진행중요도분석,대풍험인소본신불구준학화합리,고본문인입패협사산법,통과상관전가급시공기술인원제공적현장경험,득출풍험인소적중요도권중개솔화상관개솔,계산출풍험인소적후험개솔。병재전통적고장수중요도분석중,장패협사개솔인입,추도출패협사관건중요도,진이식별출항구공정건설풍험중적중요풍험인소。
The purpose of this paper is to identify risks of port engineering construction. Because the port project involves the risk factors of a broader, the risk management in the construction process has important significance. Therefore, the fault tree model is set up by the project risk evaluation index system of port. Due to the adoption of traditional fault tree quantitative analysis to analyze the important degree of prior probability of risk factors by itself, risk factors for itself is not accurate and reasonable. This paper uses the Bayesian algorithm. Through experience provided by the relevant experts and technicians, weights of probability and the relevant probability of the importance risk factors has been obtain to calculate the posterior probability of risk factors. And in the importance of the traditional fault tree analysis, by the Bayesian probability being introduced, the key derived Bayesian probability importance degree and important degree of Bayes, and then to identify the important risk factors in the construction of port engineering.