文摘
Building energy awareness and providing feedback on energy use is a vital component in transforming the behavior of individuals and communities towards a more efficient use of electric power. An enormous amount of energy consumption data has been collected over the years. In response to a growing demand for energy at multiple scales,there is a need to explore the possibility of better ways to use big data,discover meaningful patterns for autonomous energy saving opportunities,and adapt to evolving contextual data including household characteristics,global economic,and climatic trends. In this thesis,a context-aware adaptive model is proposed for optimizing energy usage across different geographic locations with appropriate comfort and satisfaction levels. The core of the model,called reverse adaptive fuzzy clustering,is the self-adaptive methodology to intelligently discover evolving features from big data of climatic and socio-economic conditions,structural and geographical attributes of households and electricity usage,and predict present and future energy demands. Patterns are iteratively obtained from matching trends of energy consumption and expenditures in US households over the years. The behavior is used as positive feedback that is converted to self-organize parameters in the model and makes a context-aware recommendation for energy optimization. The unique self-adaptive approach was examined using two real-world datasets in terms of its learning capability from past and present consumptions,context-aware personalized recommendation to reduce present energy,and self-adaptation for future energy needs.