類神經網路在產品「研發階段」可靠度管理模式之建構與分析

Implementation and Analysis of Artificial Neural Networks into Reliability Management Models in R&D Phase

尹華誠(HUA-CHEN YIN)

指導教授 ： 陳雲岫

元智大學/工程學院/工業工程與管理學系/碩士(2001年)

摘要

可靠度是產品品質的要素，可靠度的良窳直接影響產品的品質與產業競爭力；在管理機制中致力於建立整體性的可靠度管理作業制度，依環境變遷能夠適時調整因應外在的需求，應是企業經營的發展趨勢。本論文將類神經網路之技術，建構出產品「研發階段」可靠度管理模式並驗證模式的適宜性；應用類神經網路歸納與預測的動態學習特性，進行可靠度成長預測、可靠度成長環境影響因素敏感性分析、可用度預測與可靠度機率適合性曲線模式分析等，期能更充分的提供管理決策所需資訊。
以前期可靠度與各期間可靠度誤差做為神經網路輸入變數，獲致良好的可靠度成長預測績效；另使用前一期可靠度、時間與環境溫�濕度為輸入變數，以神經網路進行敏感性分析，再以成對t檢定確認環境影響因素(溫�濕度)對可靠度成長是否具有影響性；類神經網路具有良好曲線函數逼近的適用性，針對次系統失效時間間隔的發生機率，類神經網路所建立的可靠度機率適合性曲線模式所估算的機率，相較於機率分佈模式估計的機率，具有較低的誤差值。

關鍵字

可靠度管理模式 ； 類神經網路 ； 可靠度成長 ； 敏感性分析 ； 適合性曲線

並列摘要

Reliability is the essential of product quality. It directly impacts the competition of industry. A Reliability management system should be available to the rapid changes of business. This article uses artificial neural networks to establish the reliability management models for “R&D”, and verifies the validity of these models. That’s because the neural network with the characteristics of generalization and prediction forecast the reliability growth and availability. If can also analyze the sensitivity of the environmental factors of the reliability growth and the curve fitting model. Finally, we provides the information for decision making.
Use the differences of the reliabilities between the adjacent time intervals and the previous reliabilities as the input of the neural network to obtain better performance of the forecasting in reliability growth. To following, use the reliability of the right previous one, along with its time, and the humidity/temperature to have the sensitivity analysis, and then use the paired t-test to confirm the environmental factors for the reliability growth. Artificial neural network is excellent in curve fitting. It’s better in evaluating probability by curve fitting than conventional statistical models.

並列關鍵字

reliability management models ； artificial neural network ； reliability growth ； sensitivity analysis ； curve fitting

參考文獻

8.涂宏泰、高小茜、鄧維苓(1999)，應用Duane成長模式於武器操作資料可靠度評估，中華民國第三屆可靠度與維護度技術研討會論文集，397-401。

14.劉亭宜(2000)。GRNN在晶圓製造良率學習模式之建構與分析。元智大學工業工程與管理研究所碩士論文。

15.Bode, J.(1998). Decision support with neural networks in the management of research and development: Concepts and application to cost estimation. Information & Management, 34, 33-40.

16.Donovan, J. and E. Murphy (2000). A new reliability growth model: its mathematical comparison to the Duane model. Microelectronics Reliability, 40, 533-539.

17.Frost, F. and V. Karri (1999). Performance comparison of BP and GRNN models of the neural network paradigm using a practical industrial application. IEEE, 1069-1074.