Energy-Efficient and Mobility-Aware Device Selection for Task Offloading in Blockchain-Enabled IoT

Blockchain is a decentralized ledger system that securely records transactions across multiple nodes. A key challenge in blockchain networks is forking, where the transaction history diverges due to protocol changes, leading to distinct chains. In Blockchain-enabled IoT networks, forking disrupts consistency and increases latency and energy consumption. To address this issue, we propose a mobility-aware federated learning framework that optimizes the selection of IoT devices and miners to minimize forking events. Specifically, we first employed Markov Predictor to predict the future location of the devices. Further, we select the subset of devices based on their predicted locations and eliminate the devices likely to be involved in the forking event. Finally, the task of eliminating devices is reassigned to underutilized devices. Experimental results show that our Markov-based predictor achieves a 63% prediction accuracy for four to 25 devices.