Energy Optimization in WSN using Hybrid Fuzzy Q-Learning Framework for Sustainable IoT

Wireless Sensor Networks (WSNs) are important for sustainable Internet of Things (IoT) applications, but limited battery capacity, hotspot formation near the base station, and unreliable multi-hop forwarding reduce network lifetime. This paper presents a Hybrid Fuzzy Q-Learning framework for Sustainable IoT-enabled WSNs (HFQL-SIoT). The framework extends fuzzy unequal clustering and Q-learning routing by combining residual energy, base-station distance, neighbor density, node centrality, link quality, traffic load, trust score, and energy-harvesting capability for cluster-head selection. For inter-cluster routing, a multi-objective Q-learning reward function considers residual energy, distance progress, link reliability, congestion, queue delay, relay load, packet delivery ratio, trust value, and energy variance. The proposed model is evaluated with 100, 200, and 400 nodes in a 200 m x 200 m field under two base-station placements. Compared with EEUC, CHEF, and FQ-UCR, the proposed HFQL-SIoT achieves higher residual energy retention, longer stability period, improved throughput, and better packet delivery behavior. The results indicate that the integration of fuzzy decision making with reinforcement learning provides a practical energy-aware routing solution for reliable and sustainable IoT-enabled WSN deployments.