以夾擠反應曲面法構建複雜網路之近似可靠度函數

A Squeeze Response Surface Methodology to Construct Approximate Reliability Function of Complex Networks

林建興(Chien-Hsing Lin)

指導教授 ： 葉維彰

國立清華大學/工學院/工業工程與工程管理學系/碩士(2007年)

摘要

網路可靠度是管理科學研究領域裡其中一個重要的決策支援資訊。在系統設計及分析過程當中，後備零件與可靠度配置問題影響了系統未來之績效。在尋找最佳系統設置時，應考量資源受限及績效最佳化之消長關係。本論文提出了結合細胞自動機蒙地卡羅模擬法與反應曲面法之整合研究法。在模擬複雜系統網路可靠度過程中，由於總體可靠度反應值為模擬試驗之成功機率期望值，因此在逐次模擬試驗中亦可擷取出最小切割或最小路徑之資訊，透過將反應值區分為的最小切割結構函數描述之解析部分及反應曲面函數描述之模擬誤差部份以壓縮誤差區間，則此複雜系統之進似可靠度函數即由此兩部分結合而成。此一演算法可有效將複雜系統網路可靠度函數應用於各式可靠度最佳化問題當中。從以標竿問題為基礎的比較中可得知，估計解與真實可靠度解之誤差百分比在0.2%以下。由此可知，本論文所提出之研究法具有節省時間，運算過程精簡和準確預估的優點。

關鍵字

網路可靠度 ； 反應曲面法 ； 細胞自動機 ； 最小路徑/切割 ； 蒙地卡羅模擬

並列摘要

Network reliability is one of the most useful decision support information in management science. At the beginning of system design and analysis process, component redundant and reliability allocation problems affect the system performance related to availability and expect life. Searching for the optimal system configuration should consider the tradeoff between system performance and restricted resources. This thesis proposes the Squeeze Response Surface Methodology that integrate CA based MCS and RSM approach. Due to separate the simulation response into analytical and stochastic part, the range of error is squeezed. It can efficiently fit versatile models formulation that concerns to reliability optimization. From the comparison of the benchmark example, it shows that the percentage error of the purposed estimate and the exact reliability is under 0.2%. The purposed approach has the advantage of saving time, economical computation, and accurate prediction.

並列關鍵字

Network Reliability ； Response Surface Methodology (RSM) ； Cellular Automata (CA) ； MP/MC ； Monte Carlo Simulation (MCS)

參考文獻

Allan RN, Rondiris IL, Fryer DM. 1981. An efficient computational technique for evaluating the cut-tie sets and sets and common-cause failures of complex systems. Ieee Transactions on Reliability 30(2):101-109.

Box GEP, Behnken DW. 1960. Some new three level designs for the study of quantitative variables. Technometrics 2(4):455-475.

Bulka D, Dugan JB. 1994. Network s-t reliability bounds using a 2-dimensional reliability polynomial. Ieee Transactions on Reliability 43(1):39-45.

Chatelet E, Dutuit Y, Rauzy A, Bouhoufani T. 1999. An optimized procedure to generate sums of disjoint products. Reliability Engineering & System Safety 65(3):289-294.

Chen A, Yang H, Lo HK, Tang WH. 2002. Capacity reliability of a road network: an assessment methodology and numerical results. Transportation Research Part B-Methodological 36(3):225-252.