現今大多數產品可以在無故障的情況下使用很長的時間，為了能在短期內獲得足夠的產品失效訊息，可靠度領域常會使用加速壽命試驗加快試驗的速度，以有效觀測產品壽命，並推估產品在正常使用條件下的壽命，是提升產品的品質或是訂定產品使用年限的重要依據。因實驗成本及執行實驗的時間是有限的，因此如何在給定的時間內設計較有效率的實驗規劃以獲得更多產品壽命資訊是一重要的研究問題。多數文獻使用給定固定的規劃參數（planning value）設計一個局部的最佳設計（locally optimum design），然而真實參數在實驗前是未知的，因此，若給定較不準確的規劃參數作實驗設計，可能會得到較不可靠的壽命資訊。另一方面，要獲得加速壽命試驗的最佳設計通常取決於失效模型假設，例如常用的Lognormal 分配或Weibull 分配，但實務上在設計實驗的階段對確切的失效模型並不完全清楚，如果模型使用錯誤時，將會影響對壽命估計的精確度。因此，本研究將模型設定及模型參數的範圍作為考量，建構一個較穩健的大中取小設計（minimax design）。本研究透過粒子群演算法（particle swarm optimization, PSO） 來尋找可靠度加速壽命試驗的穩健設計，並驗證了PSO 尋找加速壽命試驗之大中取小設計的可行性，研究結果也顯示PSO使得我們在搜尋minimax設計時，不僅可以不用預先給定特定參數，僅需給定樣本失效機率之範圍，也可以在不同壽命模型下尋找更加穩健的設計，且搜尋到的minimax 設計比局部最佳設計相對更穩健，將更符合實務上的應用。

Because of the limitation within time constraints and experimental cost, how to plan an efficient experiment so as to obtain more lifetime information of products is an important research issue. Many literatures proposed strategies to plan a locally optimum design of an accelerated life test (ALT) under the limitation of pre-fixing the planning values of parameters. However, the true parameters are unknown before the experiment. Therefore, the experimental design under the less accurate planning parameters may obtain unreliable lifetime information. On the other hand, the optimum design for ALT usually depends on the model assumption, such as lognormal or Weibull model, but the wrong use of the model sometimes affect the accuracy of life estimation. Thus, the purpose of our study is to construct a robust minimax design that take models and parameters into consideration. To find a minimax design for ALT, we adopt particle swarm optimization (PSO) techniques. In the study, we verified the feasibility of finding minimax design via PSO. The result shows that we can search the minimax design as long as we specify the range of sample failure probability and provide candidate models. Finally, compared to the locally optimum design, the minimax design is more robust and more practical.

[1] Bahmanikashkooli, A., Zare, M., Safarpour, B., & Safarpour, M. (2014). Application of particle swarm optimization algorithm for computing critical depth of horseshoe cross section tunnel. APCBEE Procedia, 9, 207 211.
[2] Chen, P.Y., Chen, R.B., Tung, H.C., & Wong, W. K. (2017). Standardized maximim d-optimal designs for enzyme kinetic inhibition models. Chemometrics and Intelligent Laboratory Systems, 169, 79 86.
[3] Chen, R., Chang, S., Wang, W., Tung, H., & Wong, W. (2015). Minimax optimal designs via particle swarm optimization methods. Statistics and Computing, 25(5), 975–988.
[4] Ginebra, J., & Sen, A. (1998). Minimax approach to accelerated life tests. IEEE transactions on reliability, 47(3), 261–267.
[5] Hong, Y., Ma, H., & Meeker, W. Q. (2010). A tool for evaluating time-varying-stress accelerated life test plans with log-location-scale distributions. IEEE Transactions on Reliability, 59(4), 620627.
[6] King, C. B., Hong, Y., Dehart, S. P., Defeo, P. A., & Pan, R. (2016). Planning fatigue tests for polymer composites. Journal of Quality Technology, 48(3), 227245.
[7] Lee, I.C., Hong, Y., Tseng, S.T., & Dasgupta, T. (2018). Sequential bayesian design for accelerated life tests. Technometrics, 60(4), 472483.
[8] Meeker, W., & Escobar, L. (1998). Statistical methods for reliability data. Wiley.
[9] Meeker, W., & Hahn, G. (1985). How to plan an accelerated life test: Some practical guidelines. American Society for Quality Control, Statistics Division.
[10] Meeker, W. Q. (1984). A comparison of accelerated life test plans for weibull and lognormal distributions and type i censoring. Technometrics, 26(2), 157–171.
[11] Nelson, W. (1990). Accelerated testing: Statistical models, test plans, and data analysis. Wiley.
[12] Pascual, F., & Montepiedra, G. (2002). On minimax designs when there are two candidate models. Journal of Statistical Computation and Simulation, 72(11), 841862.
[13] Pascual, F. G., & Montepiedra, G. (2003). Model-robust test plans with applications in accelerated life testing. Technometrics, 45(1), 4757.
[14] Sitter, R. R. (1992). Robust designs for binary data. Biometrics, 48(4), 11451155.
[15] Xu, A., & Tang, Y. (2015). A bayesian method for planning accelerated life testing. IEEE Transactions on Reliability, 64(4), 13831392.