Using Ann with Geodesic Acceleration to Maximize Smart Energy Management

Document Type


Publication Date


Publication Title

Proceedings of the SWDSI


Home energy optimization is an increasing research interest as smart technologies in appliances and other home devices are replacing traditional items, particularly as manufacturers move to produce appliances and devices that work in conjunction with the Internet. Home energy optimization has the potential to reduce the use of energy through “smart energy management” of appliance demand for energy. Information and communications technologies (ICTs) help achieve energy savings with the goal of reducing greenhouse gas emissions and attaining effective environmental protection in several contexts including electricity generation and distribution. This “smart energy management” is utilized at the residential customer level through “smart homes.” This paper compares two artificial neural networks (ANN) used to as a support for a home energy management (HEM) system based on Bluetooth low energy, called BluHEMS. The purpose of the algorithms is to optimize energy use in a typical residential home. The first ANN uses the Levenberg-Marquardt algorithm and the second uses the Levenberg-Marquardt algorithm enhanced by a second order correction known as the geodesic acceleration.