灰色理論於美國能源消耗量預測之應用

Abstract

本研究目的要用灰色預測模型解釋並預測美國各類能源消耗量之趨勢，美國目前為全球第一大經濟體，亦為第一大能源消耗體，自從工業革命後，能源的使用越來越多，同時也帶來一些嚴重的問題，例如全球污染和全球暖化對目前全世界有潛在嚴重的風險，另外由於經濟活動，如製造業和運輸業的密集，能源效率、依賴、安全和價格等的問題也令人關注，了解未來各類能源的使用趨勢，有助市場擬定能源發展及管理計劃的關鍵。因而本文希望藉由美國做為樣本進而瞭解各類能源使用狀況與趨勢，將其研究報告作為開發中國家之樣本，以提供日後能源發展及管理方向。

本研究以GM(1,1)灰色系統理論，藉由搜集美國各類能源消耗量資訊，得知灰色系統理論在各類能源的預測均有不錯的結果。研究結果發現美國石化能源耗用與核能耗用量近年有下降趨勢，尤其2009年金融風暴後，美國能源總耗量大幅下降，理由可能是美國近年經濟衰退，且將重工業移轉至其他開發中國家，而再生能源使用量些微增加，能源耗用的趨勢顯示再生能源部分取代傳統石化能源，模型精確度分析發現，GM(1,1)灰色系統理論模型能準確預測能源的耗用，不管是石化能源或是核能預測誤差率小於10％，顯示本論文在能源耗用的建模上重大的貢獻。

本研究的管理意涵為，雖然目前傳統石化能源的使用比率依然遠高於其他種類能源，然隨著全球節能省碳與反核意識日漸高漲，再生能源技術應用滲透率（滲透率為當前市場需求除以潛在市場需求，指的是在被調查的總樣本中，某一樣本被使用的比例）也逐步提昇，善用本模型規劃能源政策，經由控制將能使能源配置適合在人類社會環境中發展，成為發達社會的關鍵先決條件。

This study uses grey forecasting models GM(1,1) to explain and predict the trend of energy consumption in the United States. U.S. is the largest economy and also is the largest energy consumer country. We attempt to understand energy consumption status in U.S., so we collect energy consumption of United States as our research sample. This investigation will provide people the direction of energy development and management. Renewable energy technology has gradually improved and grown with the consciousness of global people for antinuclear policy and carbon reduction on environment saving, although the usage of traditional fossil fuel energy is still much more than that of the other types of energy and fossil fuel is currently the most important Earth's energy. Since the industrial revolution, U.S. use more and more fossil fuels, causing such serious problems as global warming, manufacturing-intensive, and energy safety problems. Therefore, to understand the trends of the U.S. energy usage helps the U.S. and the worldwide to develop energy development and management plan. This study utilizes the GM (1,1) grey system theory to model and forecast the energy consumption. We collect all kinds of energy usage information in the United States. The results show that the consumption of renewable energy has a upward trend, while the fossil fuels has a downward trend. It suggests that renewable energy partially replace traditional fossil fuels. To our expectation, nuclear energy contrarily decreased rather than increased in a wave of anti-nuclear. Regarding the results of forecast ability, our grey model predicts different forms of energy accurately. The mean absolute percentage error is smaller than 10% for GM(1,1) model in predicting the trends of fossil fuels and nuclear energy. This suggests that the GM(1,1) performs well in predicting the energy consumption in U.S. The contribution of this investigation is to provide a suitable tools to predicting and manage energy allocation all over the world.