Decentralized Federated Learning (DFL) is recognized as a key paradigm for training models in resource-constrained, privacy-sensitive Internet of Things (IoT) environments. However, its real-world deployment is hindered by device heterogeneity, limited resources, and unpredictable node trustworthiness. To address these challenges, an innovative framework, namely Reinforcement Learning-driven Utility-based Decentralized Hierarchical Federated Learning (RL-UDHFL), is proposed, in which Reinforcement Learning (RL) is leveraged for adaptive optimization across three tiers: edge, coordination, and global aggregation. At the edge, participants are selected through an RL-Driven Participant Selection mechanism (RL-AUDPS), based on a utility function that accounts for computational resources, energy, data quality, and reputation. At the coordination level, self-tuning adaptive clustering is applied and a trust-aware gossip protocol is employed to enable robust inter-cluster communication. At the global level, reputation-based weighting is utilized and on-the-fly anomaly detection is performed to ensure model integrity. Through extensive simulations, it is demonstrated that RL-UDHFL achieves a model accuracy of 98%, surpassing hierarchical benchmarks such as HAFedRL (93.5%) and T-FedHA (92%). This superior performance is attributed to the framework’s capability to balance high accuracy, efficient resource utilization, and system reliability, thereby providing a scalable and robust blueprint for deploying sustainable and trustworthy learning systems in complex IoT applications.

The Internet of Things (IoT) comprises a vast number of interconnected devices that generate and share enormous amounts of data. Traditional machine learning approaches , which rely on the exchange of raw data, are impractical for real-world applications with extremely high data volumes due to challenges such as energy constraints and node mobility. To mitigate these overheads in IoT, Federated Learning (FL) can be employed, decentralizing the learning process to various devices without the need for centralized data collection or sharing. In this paper, we propose a new energy-efficient decentralized federated learning framework aimed at reducing energy consumption in mobile IoT. This framework utilizes a Master/Slave clustering method and a dynamic sleep/wake-up strategy, ensuring that the Base Station (BS) does not interfere with the aggregation of learning models and only supervises the clustering process . To rigorously evaluate the results of the proposed approach, we initially present a Linear Programming (LP) mathematical model designed to optimize energy consumption costs. Simulation results demonstrate that the proposed scheme improves energy consumption by up to 52 % compared to the star scheme and 41 % compared to the hierarchical method. Additionally, the proposed approach achieves a high accuracy performance of approximately 98 %, significantly surpassing standard schemes. These quantitative results highlight the effectiveness of our approach in optimizing energy use and enhancing model performance in mobile IoT environments.