在現今，產品大多具有高可靠度，生產者為了獲得產品的可靠度訊息，如產品的壽命第q百分位數，需要對產品進行實驗進而推估產品可靠度資訊。加速衰變試驗 (accelerated degradation test, ADT) 則是常被採用的一種分析工具。為了規劃有效率的加速衰變試驗，尤其是在給定總測試樣本下，如何配置樣本數至加速變數 (如溫度、壓力) 不同的應力水準上，是本研究所要探討的問題。本研究之衰變模型的建構是使用Tweedie加速衰變模型，其包含Wiener、Compound Poisson、Gamma、Inverse Gaussian等衰變模型。在D-optimality和V-optimality準則下，計算Tweedie衰變模型之費雪資訊矩陣 (Fisher information matrix) 時，部分元素沒有封閉解，而文獻上所提出之方法皆使用近似的分配去推導及分析，且未探討此近似分配對樣本數配置之影響。本研究透過數值計算的方法，計算出沒有封閉解的部分元素，並針對兩應力水準的ADT問題，在D-optimality和V-optimality準則下，討論近似方法與數值方法之最適樣本配置比例的差異。第一部分的研究是建立於固定特定模型(如Gamma process)的狀況，第二部分的研究則是在估計適合模型的狀況下去探討兩應力水準ADT問題，探討近似方法與數值方法的最適樣本配置比例的差異。最後，使用實際資料去探討最適配置問題，並在各準則下，使用相對效率 (relative efficiency) 比較最適配置與非最適配置的差異。

The accelerated degradation test (ADT) is widely used to assess the lifetime information of highly reliable products with quality characteristics. To design an efficient ADT plan, the optimum ADT allocation problem (including the determinations of stress levels, total testing times, and sample size allocations) is worthy to be discussed. This study assumes that degradation path follows Tweedie degradation model. Because of the problem that probability density function (pdf) of Tweedie degradation model does not have closed form in some cases, several researchers used the saddlepoint approximation method to approximate the probability density function. In this study, we will use numerical method to calculate the exact pdf of Tweedie degradation model. The purpose of this study is divided into two parts, the first part is established under fixed degradation model, and the second part is established under non-fixed degradation model. Under D-optimality and V-optimality criteria, we also compare the performance of two-level optimum allocation rules obtained by exact and approximated pdf. For real data, we discuss the two-level optimum allocation problem, and use relative efficiency to compare the optimum allocation rules obtained by exact and approximated pdf, respectively. The results show that the two-level optimum allocation obtained by the approximated pdf is quite efficient.

Birnbaum, Z. W., and Saunders, S. C. (1969), "A new family of life distributions," Journal of applied probability, 6(2), 319-327.
Chen, Z., Xia, T., Li, Y., and Pan, E. (2019), "Tweedie exponential dispersion processes for degradation modeling, prognostic, and accelerated degradation test planning," IEEE Transactions on Reliability, 69(3), 887-902.
Duan, F., and Wang, G. (2019), "Optimal design for constant-stress accelerated degradation test based on gamma process," Communications in Statistics-Theory and Methods, 48(9), 2229-2253.
Dunn, P. K., and Smyth, G. K. (2008), "Evaluation of Tweedie exponential dispersion model densities by Fourier inversion," Statistics and Computing, 18(1), 73-86.
Jorgensen, B. (1997), The theory of dispersion models, CRC Press.
Lim, H. (2015), "Optimum accelerated degradation tests for the gamma degradation process case under the constraint of total cost," Entropy, 17(5), 2556-2572.
Lim, H., and Yum, B. J. (2011), "Optimal design of accelerated degradation tests based on Wiener process models," Journal of Applied Statistics, 38(2), 309-325.
Meeker, W. Q., and Escobar, L. A. (2014), Statistical methods for reliability data, John Wiley & Sons.
Tseng, S. T., and Lee, I. C. (2016), "Optimum allocation rule for accelerated degradation tests with a class of exponential dispersion degradation models," Technometrics, 58(2), 244-254.
Wang, X., and Xu, D. (2010), "An inverse Gaussian process model for degradation data," Technometrics, 52(2), 188-197.
Yang, G. (2007), Life cycle reliability engineering, John Wiley & Sons.
Ye, Z. S., Chen, L. P., Tang, L. C., and Xie, M. (2014), "Accelerated degradation test planning using the inverse Gaussian process," IEEE Transactions on Reliability, 63(3), 750-763.