FEMT-FL: A novel flexible energy management technique using federated learning for energy management in IoT-based distributed green computing systems

Abstract

Green Computing systems bank on sustainable energy sources for service management and request processing. The Internet of Things (IoT) assimilates such computing systems for resource sharing and service distribution. For providing optimized energy balancing and effective utilization of available energy, this research paper proposes a novel flexible energy management technique using federated learning i.e., (FEMT-FT) for reliable energy management between the computing IoT nodes. The equivalent energy management process verifies the energy availability for request processing whereas the reliable energy part identifies the terminating interval to replace the communicating devices. For this purpose, the computing devices' draining and required energy levels are identified in all the request processing and service disseminating instances. The learning trains different energy balancing models that achieve a better service dissemination ratio. In this method, interval termination and new interval allocations are continuous to maximize service dissemination and novel request processing. For the varying requests, the proposed method achieves 8.92 %, 12.31 %, and 8.55 % high service dissemination, energy conservation, and request processing rate respectively.