- 學位論文
永續環境願景：發掘人工智慧的潛力以建構智慧生態系統
A Sustainable Environment Vision: Harnessing the Potential of Artificial Intelligence towards an Innovative Ecosystem
若您是本文的作者，可授權文章由華藝線上圖書館中協助推廣。
摘要
台灣的水、能源、材料和勞動資源受到氣候變遷的影響，因此，確保環境永續性對於未來世代的福祉至關重要。本研究提出了一種全面應對綜合性環境系統的方法，需要在宏觀、中觀和微觀層面實施政策變革。為追求與自然環境的和諧共生中，本研究重視環境管理、智慧農業的實踐，並積極探索創新的菌絲體策略。本研究所提到的環境系統中，包括四個關鍵方面：地下水位、PM2.5空氣污染嚴重程度、氣候等問題之預測及提昇生成牛樟芝的效益。在地下水位預測中，混合的人工智慧(AI)模型在R2和RMSE指標方面均超過基準，其高精確度使得其成為決策者潛在的參考依據，以進行高效的地下水資源規劃。對於PM2.5預測，混合AI模型由於額外的ACT輸入而顯著提升性能，對於準確的區域預測至關重要。該混合AI模型對於公眾意識和減少污染的政策實施是不可缺少的工具。 混合AI模型在基於中央氣象局(CWB)數據生成精確氣象預測方面表現出色，更減少對物聯網設備的依賴。這些氣象預測提供於農民將有助於溫室微氣候預測和智慧溫室的建構。該模型萃取高維數據的特徵能力强大，使其能夠準確預測農試所(TARI)和伸港(Shengang)溫室的微氣候趨勢的變化。此外，混合AI模型有助於提取顯著特徵，可靠地估計菌絲體產量並實現條件的人工智慧優化，展示了提高生產的巨大潛力，成功減少了75％的時間消耗(相比於基準)。 透過協同整治宏觀、中觀和微觀的環境系統，使得台灣更有潛力進一步朝向永續發展目標（SDGs），有效地對抗資源匱乏的威脅。此方法適用於應對眼前的環境挑戰，為未來奠定永續社會的基礎。
並列摘要
Ensuring a sustainable environment is paramount for the well-being of future generations, particularly in the face of climate-induced challenges impacting Taiwan's resources—water, energy, materials, and labor. This study advocates for a holistic approach, necessitating policy shifts at macro, meso, and micro levels, encompassing environmental management, intelligent agriculture practices, and innovative mycelia strategies—all geared towards achieving harmony with the environment. The environmental system under examination comprises four critical facets: forecasting groundwater levels, predicting PM2.5 air pollution levels, delivering climate forecasts, and enhanced mycelia yield efficiency. In groundwater level forecasting, the proposed hybrid AI model surpasses benchmarks in both R2 and RMSE metrics, providing decision-makers with robust tools for efficient groundwater resource planning. For PM2.5 level forecasting, the hybrid AI model's notable performance improvement, attributed to additional ACT inputs, is pivotal for accurate regional forecasts crucial in public awareness and policy implementation for pollution reduction. The hybrid AI model excels in generating precise climate forecasts based on Central Weather Bureau (CWB) data, reducing reliance on IoT devices. These climate forecasts are extended to farmers, aiding in greenhouse microclimate predictions and the construction of intelligent greenhouses. The model's adept feature extraction from high-dimensional datasets enables accurate forecasts of greenhouse microclimate trends for TARI and Shengang greenhouses. Furthermore, integrating AI facilitates the extraction of salient features for reliable mycelia yield estimation and AI-driven optimization of conditions, showcasing significant potential for production enhancement and a notable 75% reduction in time consumption compared to conventional methods. In synergizing macroscale, mesoscale, and microscale environmental systems, Taiwan takes strides toward realizing Sustainable Development Goals (SDGs), effectively countering the specter of resource scarcity. This comprehensive approach addresses immediate environmental challenges and sets the stage for a sustainable and resilient future.
參考文獻
Adams, S. R., Cockshull, K. E., & Cave, C. R. J. (2001). Effect of temperature on the growth and development of tomato fruits. Annals of botany, 88(5), pp. 869-877. https://doi.org/10.1006/anbo.2001.1524
Al-Yahyai, S., Charabi, Y., & Gastli, A. (2010). Review of the use of numerical weather prediction (NWP) models for wind energy assessment. Renewable and Sustainable Energy Reviews, 14(9), 3192-3198. https://doi.org/10.1016/j.rser.2010.07.001
Andoko, E., Liu, W. Y., Zeng, H. J., & Sjoblom, A. (2020). Review of Taiwan’s food security strategy. Food and Fertilizer Technology Center for the Asian and Pacific Region (FFTC-AP)(website).
Akkaya, E. (2016). ANFIS based prediction model for biomass heating value using proximate analysis components. Fuel, 180, 687-693.
Arias, A., Feijoo, G., & Moreira, M. T. (2023). How could Artificial Intelligence be used to increase the potential of biorefineries in the near future? A review. Environmental Technology & Innovation, 32, 103277.