葉藝勇



摘要為了對廣東省的能源需求進行準確的預測,首先分析了影響廣東省能源需求的各種因素,構建了預測指標體系.在此基礎上,針對能源系統非線性等復雜系統特征,結合粒子群算法和BP神經網絡的優點,構建了改進的PSOBP神經網絡的預測模型,并通過主成分分析法對指標體系進行數據降維,以降低神經網絡的規模和復雜程度.以廣東省1985-2013年的能源需求數據進行模擬與仿真,并對2014-2018年的能源需求量進行預測,理論分析和實證研究表明,該方法能夠很好的反映廣東省能源需求的特征,預測結果較為準確合理.
關鍵詞能源需求預測,粒子群算法,BP神經網絡,主成分分析法
中圖分類號F201 文獻標識碼A
Construction of Energy Demand Forecasting
Model and Empirical Analysis of Guangdong Province
YE Yiyong
(College of Economics & Management Wuyi University Jiangmen, Guangdong 529020,China)
AbstractIn order to make accurate forecast for energy demand of Guangdong province, this paper analyzed the various factors which impact on energy demand of Guangdong province, and constructed the predict index system. On this basis, according to the nonlinear characteristics of the energy system, combined with the advantages of particle swarm optimization algorithm and BP neural network, a prediction model was constructed based on PSOBP neural network. And the method of principal component analysis was used to reduce the dimensions of the prediction index system in order to reduce the size and complexity of the neural network. Then, this paper simulated the energy demand data of Guangdong province from 1985 to 2013, and carried on the forecast energy demand of Guangdong province during 2014 to 2018. The theoretical analysis and empirical study show that this method can reflect the characteristics of energy demand of Guangdong province, and the predicted result is more accurate and reasonable.
Key words forecasting of energy demand, PSO, BP neural network, PCA
1引言
隨著社會經濟的快速發展,各行業對能源的需求大幅度增加.據統計,廣東省2000年的能源消耗量是7 983萬噸標準煤,2013年的能源消耗量上升到25 645萬噸標準煤,是2000年消耗量的3.2倍,其中一次能源消費90%依賴省外,二次能源消費中的電力消費有10%也是依賴省外,據估算,未來10年這個比例將達到30%左右.經濟快速發展所帶來的巨大能源需求與供給不足之間的矛盾越來越嚴重,能源短缺已成為制約廣東省經濟持續發展的關鍵問題,如果不采取有效的措施,將會延緩廣東省產業結構的轉型升級優化,乃至影響全省經濟的穩步增長.系統地分析廣東省能源需求的影響因素,準確地預測廣東省未來能源需求的數量,進而制定科學合理的能源發展……