TAS2MOO-ICRL: A Traffic and Security-Aware Multi-Objective Optimization Framework Using Inverse Constrained Reinforcement Learning for Energy Management in Wireless Sensor Networks
Wireless Sensor Networks (WSNs) operating in mission-critical environments require statistically guaranteed trade-offs among energy usage, communication latency, and security strength. This paper introduces TAS2MOO-ICRL, a Traffic and Security-Aware Multi-Objective Optimization framework using Inverse Constrained Reinforcement Learning to statistically model and trade these competing objectives. Our suggested approach uses a multi-objective RL model with a dynamic reward-shaping function that combines energy consumption metrics, end-to-end delay statistics, and intrusion risk probabilities to maximize efficiency through strategic node sleep scheduling. Traffic load is predicted with simple predictive time-series models, allowing for a proactive adjustment of node activity in reaction to statistically significant load changes and indicators of security threats. Simulation analyses reveal that TAS2MOO-ICRL achieves a statistically significant 30% reduction in average node energy consumption (p < 0.01), a 50% decrease in average packet latency (p < 0.01), and a 93.7% intrusion detection rate with a false positive rate below 7%, outperforming baseline methods. These results confirm TAS2MOO-ICRL as a statistically robust framework for improving WSN operational longevity whilst ensuring an efficient communication standard and a high resiliency against security threats.