地層滲透率是指流體（石油、天然氣、水及二氧化碳等）在地底下的流動能力，對於油氣生產、地熱資源開採及二氧化碳地質封存都極為重要。獲得地層滲透率的傳統方法主要為：岩心分析 (core analysis) 與井壓測試分析 (well test analysis)。
岩心分析可以得到特定深度點的滲透率數值，然而取得岩心樣本的過程非常昂貴且困難，其準確度與取岩心的過程及樣本完整性皆有關。因此，在大部分的開發場域中，岩心樣本的滲透率資料量都非常稀少。井壓測試分析藉由計算井壓與地下流體從井中流出的流量變化關係，來求得地層滲透率。井壓測試分析所得到的滲透率為測試地層穿孔區間的平均值，因此無法得知地層滲透率在垂直方向的變化趨勢。
與岩心資料不同，幾乎所有的井都有種類豐富且連續的井測資料，利用井測資料計算相關的地層參數也已是一項成熟的技術。傳統計算時須使用各種經驗公式，但這些公式中有許多參數的因子只適用於特定岩性或區域。因此，利用類神經網路來尋找井測資料與岩心資料之間的潛在關係，並藉此得到地層滲透率，是近年來發展出的新方法。本研究以機器學習的方式預測地層滲透率，目的在節省取得岩心資料成本的同時，仍能得到地層滲透率在垂直方向的變化趨勢，是其最大的優勢。
自適應差分進化演算法 (Self-adaptive Differential Evolution, SADE) 是一種具備全域搜索能力的最佳化演算方法，本研究將其應用在多層前饋式類神經網路，作為訓練模型的優化器，建立自適應差分進化演算法類神經網路 (Self-adaptive Differential Evolution Artificial Neural Network, SADE-ANN) 模型，取代傳統的倒傳遞類神經網路 (Back-propagation Artificial Neural Network, BP-ANN)。希望可以藉由SADE解決梯度下降法中，容易陷入局部最佳解無法跳脫的困境，並透過其強大的全域搜索能力，來彌補岩心樣本稀少所可能導致的類神經網路學習不足的問題。
本研究所提出的SADE，改良了原本的差分進化演算法 (Differential Evolution, DE)，在設定模型的控制參數時，可以根據每一次迭代的收斂趨勢作為學習經驗，逐漸自我適應，自動調整每一世代最適合的控制參數，減少人為調整參數的比例，以此提升最佳化的效能。並設計一個適合應用於SADE-ANN模型的交叉驗證方法，使其保有SADE全域搜索的特性，同時避免產生過度擬合的現象。最後以均方誤差 (MSE) 及決定係數 (R2) 兩項指標，來評估並比較SADE-ANN模型及BP-ANN模型的地層滲透率預測能力。
本研究所建立的SADE-ANN模型，以井測資料作為輸入值，岩心資料作為輸出值，以110筆資料作為學習樣本，便成功預測地層滲透率，克服了學習樣本稀少，導致模型訓練不足的問題。以R2作為模型效能的評估標準，BP-ANN模型為47.1%，SADE-ANN模型的預測解釋力則達到78.1%。代表SADE-ANN模型所得到的滲透率預測值具有高可信度，並且其模型效能優於傳統的BP-ANN模型。

Permeability refers to the ability of a porous rock formation to allow fluids, such as oil, natural gas, water, and carbon dioxide, to flow through it. It is important for hydrocarbon production, geothermal resource exploitation, and geological storage of carbon dioxide. Traditional methods for obtaining permeability involve core analysis and well test analysis. Both methods require significant amounts of time and money. Therefore, many research using artificial neural networks to explore the potential relationships between well log data and core data, and thereby obtain permeability values.
In this study, a new machine learning regression model, the Self-adaptive Differential Evolution Artificial Neural Netwok (SADE-ANN) model, is proposed. It improves the existing Self-adaptive Differential Evolution (SADE) algorithm and employs it as an optimizer for Multi-layer Feed Forward Neural Netwok (MFFNN). Its purpose is to address the issue of the gradient descent method easily getting trapped in local optima. By incorporating appropriate cross-validation techniques, it ensures the global search capability of SADE while preventing the occurrence of over-fitting in the Artificial Neural Netwok (ANN). This allows for its application in engineering practices with limited training samples.
The SADE-ANN model utilizes well log data and core data as learning samples, with 110 learning samples, successfully predicts permeability. Using Coefficient of Determination (R2) as the evaluation criterion, SADE-ANN model exhibits better performance than the Back-propagation ANN (BP-ANN) model, with R2 of 0.781, which means the SADE-ANN model are highly credible.

Abdual-Salam, M. E., Abdul-Kader, H. M., & Abdel-Wahed, W. F. (2010). Comparative study between differential evolution and particle swarm optimization algorithms in training of feed-forward neural network for stock price prediction. Paper presented at the 2010 the 7th international conference on informatics and systems (INFOS).
Alguliev, R. M., Aliguliyev, R. M., & Isazade, N. R. (2012). DESAMC+ DocSum: Differential evolution with self-adaptive mutation and crossover parameters for multi-document summarization. Knowledge-Based Systems, 36, 21-38.
Ali, M., & Ahn, C. W. (2014). An optimized watermarking technique based on self-adaptive DE in DWT–SVD transform domain. Signal processing, 94, 545-556.
Archer, J. S., & Wall, C. G. (2012). Petroleum engineering: principles and practice: Springer Science & Business Media.
Asquith, G., Krygowski, D., Henderson, S., & Hurley, N. (2004). Basic relationships of well log interpretation.
Asquith, G. B., Krygowski, D., & Gibson, C. R. (2004). Basic well log analysis (Vol. 16): American Association of Petroleum Geologists Tulsa.
Babatunde, O. M., Munda, J. L., & Hamam, Y. (2020). Exploring the potentials of artificial neural network trained with differential evolution for estimating global solar radiation. Energies, 13(10), 2488.
Basheer, I. A., & Hajmeer, M. (2000). Artificial neural networks: fundamentals, computing, design, and application. Journal of microbiological methods, 43(1), 3-31.
Bengio, Y., & Grandvalet, Y. (2003). No unbiased estimator of the variance of k-fold cross-validation. Advances in Neural Information Processing Systems, 16.
Bhatia, S., & Vishwakarma, V. P. (2016). Feed forward neural network optimization using self adaptive differential evolution for pattern classification. Paper presented at the 2016 IEEE International Conference on Recent Trends in Electronics, Information & Communication Technology (RTEICT).
Bradley, H. B. (1987). Petroleum engineering handbook.
Brest, J., Greiner, S., Boskovic, B., Mernik, M., & Zumer, V. (2006). Self-adapting control parameters in differential evolution: A comparative study on numerical benchmark problems. IEEE transactions on evolutionary computation, 10(6), 646-657.
Bui, N. T., & Hasegawa, H. (2015). Training artificial neural network using modification of differential evolution algorithm. International journal of Machine Learning and computing, 5(1), 1.
Chen, C. (2006). Committee-machine-based models for permeability prediction. Doctoral thesis, National Cheng Kung University of Taiwan, Taiwan,
Chin, W. W. (1998). The partial least squares approach to structural equation modeling. Modern methods for business research, 295(2), 295-336.
Cohen, J. (2016). A power primer.
Dake, L. P. (1983). Fundamentals of reservoir engineering: Elsevier.
Darwin’s, C. (1859). On the origin of species. published on, 24, 1.
Das, S., & Suganthan, P. N. (2010). Differential evolution: A survey of the state-of-the-art. IEEE transactions on evolutionary computation, 15(1), 4-31.
Deng, L.-B., Wang, S., Qiao, L.-Y., & Zhang, B.-Q. (2017). DE-RCO: rotating crossover operator with multiangle searching strategy for adaptive differential evolution. IEEE Access, 6, 2970-2983.
Do, H. (1949). The organization of behavior. New York.
Doll, H. t. (1949). The SP log: Theoretical analysis and principles of interpretation. Transactions of the AIME, 179(01), 146-185.
Elkatatny, S., Mahmoud, M., Tariq, Z., & Abdulraheem, A. (2018). New insights into the prediction of heterogeneous carbonate reservoir permeability from well logs using artificial intelligence network. Neural Computing and Applications, 30, 2673-2683.
Fanchi, J. R., & Christiansen, R. L. (2016). Introduction to petroleum engineering: John Wiley & Sons.
Fortmann-Roe, S. (2012). Understanding the bias-variance tradeoff. URL: http://scott. fortmann-roe. com/docs/BiasVariance. html (hämtad 2019-03-27).
Gämperle, R., Müller, S. D., & Koumoutsakos, P. (2002). A parameter study for differential evolution. Advances in intelligent systems, fuzzy systems, evolutionary computation, 10(10), 293-298.
Goh, A. T. (1995). Back-propagation neural networks for modeling complex systems. Artificial intelligence in engineering, 9(3), 143-151.
Harris, D. C. (2010). Quantitative chemical analysis: Macmillan.
Hecht-Nielsen, R. (1988). Applications of counterpropagation networks. Neural networks, 1(2), 131-139.
Hecht-Nielsen, R. (1990). Neurocomputing, Addison. Wesely Publishing Company. Hornik, K. Stinchcombe, M. White, H.(1989). Multilayer feedforward networks are universal approximators, Neural Networks, 2(359366), 3168-3176.
Hopfield, J. J., & Tank, D. W. (1986). Computing with neural circuits: A model. Science, 233(4764), 625-633.
Hsieh, B.-Z. (2006). 井測法-地層岩性及石油蘊藏探勘. 科學發展.
Ilonen, J., Kamarainen, J.-K., & Lampinen, J. (2003). Differential evolution training algorithm for feed-forward neural networks. Neural Processing Letters, 17, 93-105.
Jain, A. K., Mao, J., & Mohiuddin, K. M. (1996). Artificial neural networks: A tutorial. Computer, 29(3), 31-44.
Kamble, R. G., Raykar, N., & Jadhav, D. (2021). Machine learning approach to predict fatigue crack growth. Materials Today: Proceedings, 38, 2506-2511.
Lim, J.-S., & Kim, J. (2004). Reservoir porosity and permeability estimation from well logs using fuzzy logic and neural networks. Paper presented at the SPE Asia Pacific Oil and Gas Conference and Exhibition.
Malki, H., Baldwin, J., & Kwari, M. (1996). Estimating permeability by use of neural networks in thinly bedded shaly gas sands. SPE Computer Applications, 8(02), 58-62.
McCarthy, J., Minsky, M. L., Rochester, N., & Shannon, C. E. (2006). A proposal for the dartmouth summer research project on artificial intelligence, august 31, 1955. AI magazine, 27(4), 12-12.
McCord, N. M., & Illingworth, W. (1990). A practical guide to neural nets. Reading: Addison-Wesley [Material Impreso en biblioteca].
Mehrotra, K., Mohan, C. K., & Ranka, S. (1997). Elements of artificial neural networks: MIT press.
Meza, G. R., Ferragud, X. B., Saez, J. S., & Durá, H. (2017). Controller tuning with evolutionary multiobjective optimization. Switzerland: Springer.
Michalewicz, Z., & Michalewicz, Z. (1996). GAs: Why Do They Work? : Springer.
Mohaghegh, S., Arefi, R., Bilgesu, I., Ameri, S., & Rose, D. (1995). Design and development of an artificial neural network for estimation of formation permeability. SPE Computer Applications, 7(06), 151-154.
Moore, D. S., & Kirkland, S. (2007). The basic practice of statistics (Vol. 2): WH Freeman New York.
Moussa, T., Elkatatny, S., Mahmoud, M., & Abdulraheem, A. (2018). Development of new permeability formulation from well log data using artificial intelligence approaches. Journal of Energy Resources Technology, 140(7), 072903.
Ott, R. L., & Longnecker, M. T. (2015). An introduction to statistical methods and data analysis: Cengage Learning.
Piotrowski, A. P. (2014). Differential evolution algorithms applied to neural network training suffer from stagnation. Applied Soft Computing, 21, 382-406.
Price, K., Storn, R. M., & Lampinen, J. A. (2006). Differential evolution: a practical approach to global optimization: Springer Science & Business Media.
Qin, A. K., & Suganthan, P. N. (2005). Self-adaptive differential evolution algorithm for numerical optimization. Paper presented at the 2005 IEEE congress on evolutionary computation.
Reynoso-Meza, G., Blasco, X., Sanchis, J., & Martínez, M. (2014). Controller tuning using evolutionary multi-objective optimisation: current trends and applications. Control Engineering Practice, 28, 58-73.
Rider, M. H. (1986). The geological interpretation of well logs.
Rodriguez-Galiano, V., Sanchez-Castillo, M., Chica-Olmo, M., & Chica-Rivas, M. (2015). Machine learning predictive models for mineral prospectivity: An evaluation of neural networks, random forest, regression trees and support vector machines. Ore Geology Reviews, 71, 804-818.
Rodriguez, J. D., Perez, A., & Lozano, J. A. (2009). Sensitivity analysis of k-fold cross validation in prediction error estimation. IEEE transactions on pattern analysis and machine intelligence, 32(3), 569-575.
Rosenblatt, F. (1962). Principles of neurodynamics: Perceptrons and the theory of brain mechanisms (Vol. 55): Spartan books Washington, DC.
Schalkoff, R. J. (1997). Artificial neural networks: McGraw-Hill Higher Education.
Sharma, S., Sharma, S., & Athaiya, A. (2017). Activation functions in neural networks. Towards Data Sci, 6(12), 310-316.
Singh, A., Thakur, N., & Sharma, A. (2016). A review of supervised machine learning algorithms. Paper presented at the 2016 3rd international conference on computing for sustainable global development (INDIACom).
Slowik, A. (2010). Application of an adaptive differential evolution algorithm with multiple trial vectors to artificial neural network training. IEEE Transactions on Industrial Electronics, 58(8), 3160-3167.
Slowik, A., & Bialko, M. (2008). Training of artificial neural networks using differential evolution algorithm. Paper presented at the 2008 conference on human system interactions.
Soliman, O. S., & Bui, L. T. (2008). A self-adaptive strategy for controlling parameters in differential evolution. Paper presented at the 2008 IEEE Congress on Evolutionary Computation (IEEE World Congress on Computational Intelligence).
Spears, W. M., De Jong, K. A., Bäck, T., Fogel, D. B., & De Garis, H. (1993). An overview of evolutionary computation. Paper presented at the European conference on machine learning.
Storn, R. (2008). Differential evolution research–trends and open questions. In Advances in differential evolution (pp. 1-31): Springer.
Storn, R., & Price, K. (1997). Differential evolution–a simple and efficient heuristic for global optimization over continuous spaces. Journal of global optimization, 11, 341-359.
Vesterstrom, J., & Thomsen, R. (2004). A comparative study of differential evolution, particle swarm optimization, and evolutionary algorithms on numerical benchmark problems. Paper presented at the Proceedings of the 2004 congress on evolutionary computation (IEEE Cat. No. 04TH8753).
Wang, S.-C. (2003). Interdisciplinary computing in Java programming (Vol. 743): Springer Science & Business Media.
Wang, S.-C., & Wang, S.-C. (2003). Artificial neural network. Interdisciplinary computing in java programming, 81-100.
Wythoff, B. J. (1993). Backpropagation neural networks: a tutorial. Chemometrics and intelligent laboratory systems, 18(2), 115-155.
Yao, X. (1999). Evolving artificial neural networks. Proceedings of the IEEE, 87(9), 1423-1447.
Yeo, I. K., & Johnson, R. A. (2000). A new family of power transformations to improve normality or symmetry. Biometrika, 87(4), 954-959.
Zhang, G., Patuwo, B. E., & Hu, M. Y. (1998). Forecasting with artificial neural networks:: The state of the art. International journal of forecasting, 14(1), 35-62.
Zhu, Q., Lin, Q., Du, Z., Liang, Z., Wang, W., Zhu, Z., . . . Ming, Z. (2016). A novel adaptive hybrid crossover operator for multiobjective evolutionary algorithm. Information Sciences, 345, 177-198.
Zikmund, W. G., Babin, B. J., Carr, J. C., & Griffin, M. (2013). Business research methods: Cengage learning.
林豐澤. (2005a). 演化式計算上篇: 演化式演算法的三種理論模式. 智慧科技與應用統計學報, 3(1), 1-28.
林豐澤. (2005b). 演化式計算下篇: 基因演算法以及三種應用實例. 智慧科技與應用統計學報, 3(1), 29-56.