A neural network-based appliance scheduling methodology for smart homes and buildings with multiple power sources

The increased production of electrical energy from various power sources, such as solar, wind, and nuclear, allows smart buildings to be connected with multiple power sources. In an effort to conserve environment, electrical energy usage is gradually shifting towards renewable and green energy sources, such as wind, hydro, and solar. Regardless of power sources, a user demands for continuous energy supply and also desires to minimize the electricity bill. Further, in a dynamic pricing environment, the price of electricity varies throughout the day. In such a dynamic pricing environment, appliance scheduling with multiple power sources tied to a smart home/building is an important research problem. In this paper, we propose a methodology to abet green environment by prioritizing green energy sources and to minimize the electricity cost for the user. Our proposed methodology leverages a smart grid architecture which employs a greedy strategy to select the most feasible power source amongst the available power sources tied to a smart home/building. Our proposed methodology further leverages a neural network-based approach for appliance scheduling that optimizes the use of power sources in a dynamic pricing environment to minimize the total cost of electricity usage.