臺灣博碩士論文加值系統

維持穏定的電力供應是許多國家持續發展經濟建設的重要議題，為滿足台灣持續的經濟發展及環境保護，未來勢必要開發新的綠色能源發電廠，尤其是要利用離島潛在的日光及風力資源。準確的未來電力需求預測是開發新電力資源及電力系統經濟調度中的重要參考因素。本研究透過類群體演算法(CSA)建置三種電力需求預測模型，因為CSA繼承粒子群演算法(PSO)的特性，其搜尋範圍大、參數少、收斂速度快，可以快速的在問題空間中找出近似最佳解。同時CSA引進參考解和路徑重劃的概念，可以避免搜尋解困入區域最佳解之中。本研究透過有參數迴歸模型和無參數高斯迴歸模型處理台灣歷史每月的電力消費資料，有參數迴歸模型參考影響電力消費的重要指標，無參數高斯迴歸模型利用高斯函數達到逼近歷史的電力消費資料，再因電力消費的週期特性以平移取得電力需求預測，最後以強化式學習整合電力需求預測模型來找出更最佳的預測結果，實驗結果得到2%~3%的預測平均誤差百分比，也因此本研究提出的方法獲得相當優異的預測水準。

Stable electricity supply is the basis of economic development for many countries. In order to satisfy the electric power demand of Taiwan's future economic development and environmental protection, new power plants using green energy must be developed, especial the potential solar and wind energy sources in off-island areas. Accurate electricity demand forecasting plays an important role in the economic dispatch of electricity system. This study develops three demand forecasting models based on the Cyber Swarm Algorithm (CSA) which inherits the main features from the particle swarm optimization algorithm. Because of its prevailing capability such as large searching range, few parameters, and fast convergence speed, the particle swarm optimization algorithm usually can quickly find a near optimal solution in the solution space. Moreover, CSA employs the search strategies introduced in reference set and path relinking in order to escape from the barrier of local optima. This study applies parametric and non-parametric regression models to process electricity demand data. Parametric models use important indicators for predicting electricity demand, while non-parametric models use Gauss functions to approximate historical data and shift the learned functions to predict future electricity demand. Finally, we propose a reinforcement learning model to integrate parametric and non-parametric models and obtain a better prediction result.

目次
誌謝 I
摘要 II
Abstract III
目次 IV
圖目次 VI
表目次 VII
第一章 緒論 1
1.1 研究背景 1
1.2 研究動機 2
1.3 研究目的 2
1.4 研究架構 2
第二章 文獻回顧 3
2.1 電力預測問題 3
2.2 電力預測方法 7
2.3 次經驗法則 8
第三章 研究方法 9
3.1 問題定義 9
3.2 有參數迴歸模型 10
3.3 高斯函數逼近模型 15
3.4 CSA最佳化預測模型 21
3.5 強化式學習 24
第四章 實驗結果 26
4.1 效能評估設計 26
4.2 迴歸模型預測分析 29
4.3 高斯模型預測分析 32
4.4 強化式預測分析 36
第五章 結論與建議 40
參考文獻 41

圖目次
圖 1　2012-2014年每月用電消費量 11
圖 2　2012-2014年台灣每月平均溫度 11
圖 3　2001-2014年各部門佔消費電量比例 12
圖 4　2001-2014年各部門消費電量 13
圖 5　2001-2014年進出口額 13
圖 6　2001-2014年各部門電力消費成長量 14
圖 7　2001-2014每年人口、電力消費量 14
圖 8　部分電力消費量曲線圖 16
圖 9　高斯函數曲線圖 17
圖 10　高斯函數解最佳逼近示意圖 17
圖 11　平移函數預測示意圖 18
圖 12　投影平均示意圖 20
圖 13　CSA流程 23
圖 14　系統流程圖 25
圖 15　二次模型距離參數最佳化 30
圖 16　指數模型距離參數最佳化 30
圖 17　迴歸模型預測圖 31
圖 18　高斯模型距離參數最佳化 34
圖 19　高斯模型預測圖 36
圖 20　各模型收斂分析 38
圖 21　三模型強化式學習預測 39

表目次
表 1　演算法參數設定 27
表 2　指數模型與文獻比較結果 28
表 3　二次模型與文獻比較結果 28
表 4　迴歸模型變數設定 29
表 5　迴歸模型預測結果 31
表 6　高斯逼近模型變數設定 32
表 7　高斯個數參數最佳化 33
表 8　平移模型變數設定 34
表 9　高斯模型預測結果 35
表 10　有參數迴歸模型混合預測結果 37
表 11　各模型之誤差比較 38
表 12　混合預測結果 40

Askarzadeh, A. (2014). Comparison of particle swarm optimization and other metaheuristics on electricity demand estimation: A case study of Iran. Energy, 72, 484-491.
Dorigo,M.,Maniezzo,V.,&Colorni,A.(1996)Ant system:optimization by a colony of cooperating agents.Systems,Man and Cybernetics,PartB:CIEEE Transactions on,26(1),29-41.
Eberhart,R.C.,&Kennedy,J.(1995).A new optimizer using particle swarm theory.paper presented at the Proceedings of the sixth international symposium on micro machine and human science.
Esteves, G. R., Bastos, B. Q., Cyrino, F. L., Calili, R. F., & Souza, R. C. (2015). Long Term Electricity Forecast: A Systematic Review. Procedia Computer Science, 55, 549-558.
Filik, Ü. B., Gerek, Ö. N., & Kurban, M. (2011). A novel modeling approach for hourly forecasting of long-term electric energy demand. Energy conversion and management, 52(1), 199-211.
Glover,F.(1989).Tabu search-partI.ORSA Journal on computing,1(3),190-206
González-Romera, E., Jaramillo-Morán, M., & Carmona-Fernández, D. (2008). Monthly electric energy demand forecasting with neural networks and Fourier series. Energy conversion and management, 49(11), 3135-3142.
Hernandez, L., Baladron, C., Aguiar, J. M., Carro, B., Sanchez-Esguevillas, A. J., Lloret, J., & Massana, J. (2014). A survey on electric power demand forecasting: future trends in smart grids, microgrids and smart buildings. Communications Surveys & Tutorials, IEEE, 16(3), 1460-1495.
Hippert, H. S., Pedreira, C. E., & Souza, R. C. (2001). Neural networks for short-term load forecasting: A review and evaluation. Power Systems, IEEE Transactions on, 16(1), 44-55.
Holland, J. H. (1975). Adaptation in natural and artificial systems: an introductory analysis with applications to biology, control, and artificial intelligence. U Michigan Press.
Huang, C.-M., Huang, C.-J., & Wang, M.-L. (2005). A particle swarm optimization to identifying the ARMAX model for short-term load forecasting. Power Systems, IEEE Transactions on, 20(2), 1126-1133.
Huang, S.-J., & Shih, K.-R. (2003). Short-term load forecasting via ARMA model identification including non-Gaussian process considerations. Power Systems, IEEE Transactions on, 18(2), 673-679.
Jaramillo-Morán, M. A., González-Romera, E., & Carmona-Fernández, D. (2013). Monthly electric demand forecasting with neural filters. International Journal of Electrical Power & Energy Systems, 49, 253-263.
Kıran, M. S., Özceylan, E., Gündüz, M., & Paksoy, T. (2012). A novel hybrid approach based on particle swarm optimization and ant colony algorithm to forecast energy demand of Turkey. Energy conversion and management, 53(1), 75-83.
Lee, W.-J., & Hong, J. (2015). A hybrid dynamic and fuzzy time series model for mid-term power load forecasting. International Journal of Electrical Power & Energy Systems, 64, 1057-1062.
Moghram, I., & Rahman, S. (1989). Analysis and evaluation of five short-term load forecasting techniques. Power Systems, IEEE Transactions on, 4(4), 1484-1491.
Nawaz, S., Iqbal, N., & Anwar, S. (2014). Modelling electricity demand using the STAR (Smooth Transition Auto-Regressive) model in Pakistan. Energy, 78, 535-542.
Pao, H. (2009). Forecasting energy consumption in Taiwan using hybrid nonlinear models. Energy, 34(10), 1438-1446.
Wang, J., Zhu, W., Zhang, W., & Sun, D. (2009). A trend fixed on firstly and seasonal adjustment model combined with the ε-SVR for short-term forecasting of electricity demand. Energy Policy, 37(11), 4901-4909.
Yin, P. Y., Glover, F., Laguna, M., & Zhu, J. X. (2010). Cyber swarm algorithms–improving particle swarm optimization using adaptive memory strategies. European Journal of Operational Research, 201(2), 377-389.
Zhang, W. Y., Hong, W.-C., Dong, Y., Tsai, G., Sung, J.-T., & Fan, G.-f. (2012). Application of SVR with chaotic GASA algorithm in cyclic electric load forecasting. Energy, 45(1), 850-858.