Adaptive MARL-Assisted Hybrid Bat-Artificial Bee Colony Optimization for Energy-Efficient Clustering and Routing in...

Area of Science:

• Computer Science
• Electrical Engineering
• Artificial Intelligence

Background:

• Energy efficiency is critical for Internet of Things (IoT) wireless sensor networks (WSNs) due to limited node battery life and impractical replacement.
• Current clustering and routing protocols often optimize these functions separately, causing energy depletion, node failure, and reduced network reliability.

Purpose of the Study:

• To propose an integrated framework, MARL-BA-ABC, for joint cluster-head selection and routing optimization in IoT-enabled WSNs.
• To enhance network lifetime and performance by addressing limitations of existing separate optimization approaches.

Main Methods:

• Developed an Adaptive Multi-Agent Reinforcement Learning-Assisted Hybrid Bat-Artificial Bee Colony (MARL-BA-ABC) framework.
• Jointly optimized cluster-head selection and routing using parameters like residual energy, distance, traffic load, node density, and link quality.
• Formulated a multi-objective optimization model to minimize energy consumption, delay, overhead, and load imbalance, while maximizing packet delivery ratio and network lifetime.

Main Results:

• The MARL-BA-ABC framework demonstrated superior performance compared to LEACH, HEED, PSO, ACO, BA, and ABC.
• Achieved a First Node Death at 2200 rounds, residual energy of 1.32 J, packet delivery ratio of 98.1%, throughput of 410 kbps, and average end-to-end delay of 7.4 ms.

Conclusions:

• The proposed MARL-BA-ABC framework effectively enhances energy efficiency and network performance in IoT-enabled WSNs.
• Joint optimization of cluster-head selection and routing significantly improves network reliability and extends operational lifetime.