ABSTRACT
This research presents an intelligent fog computing framework
designed to enhance Quality of Service (QoS) in Internet of Things (IoT)
environments through the integration of fuzzy logic, Deep Q-Network (DQN),
and Multi-Objective Evolutionary Algorithm based on Decomposition
(MOEA/D). The system employs fuzzy logic to classify tasks by priority using
parameters such as task size, device load, and network congestion; a DQN
scheduler then learns optimal task-to-node mappings, while MOEA/D
performs multi-objective resource allocation balancing latency, reliability, and
utilization. Implemented and tested in an iFogSim2 environment, the
proposed model achieved a task classification accuracy of 96.84%, task
success rate above 96%, deadline completion ratio exceeding 96%, and more
than 10% latency reduction compared with leading baseline models. The
framework demonstrated strong scalability, adaptability, and reliability,
achieving 90.1% resource utilization and rapid learning convergence. These
findings indicate that the proposed hybrid architecture provides an effective,
self-optimizing fog-IoT system suitable for real-time applications such as smart
healthcare, urban management, and vehicular coordination, with future
extensions envisioned for energy efficiency, trust-based computation, and
automated parameter tuning.
