An improved human memory algorithm with multi-directional and chaotic approaches for global optimization and...

Area of Science:

• Computer Science
• Electrical Engineering
• Network Engineering

Background:

• Wireless Sensor Networks (WSNs) are vital for infrastructure monitoring but face energy efficiency challenges due to limited node battery life.
• Cluster Head (CH) selection is critical for WSN energy efficiency, yet traditional metaheuristic algorithms often suffer from premature convergence.
• Existing methods struggle with the dynamic and energy-sensitive nature of CH selection in WSNs, leading to suboptimal performance.

Purpose of the Study:

• To develop an advanced optimization algorithm, Adaptive Enhanced Human Memory Optimization (AEHMO), for energy-efficient CH selection in WSNs.
• To enhance the Human Memory Optimization (HMO) algorithm by integrating adaptive parameters, multi-directional mutation, dynamic drift search, and chaotic reverse learning.
• To address the limitations of traditional metaheuristic algorithms in dynamic WSN environments.

Main Methods:

• Developed Adaptive Enhanced Human Memory Optimization (AEHMO) by enhancing the Human Memory Optimization (HMO) algorithm.
• Integrated four key strategies into AEHMO: Adaptive parameters, Multi-Directional Mutation Strategy (MDMS), Dynamic Drift Search (DDS), and Chaotic Reverse-based Learning (CRL).
• Validated AEHMO on the CEC2017 benchmark suite for high-dimensional optimization and applied it to homogeneous and heterogeneous WSN scenarios.

Main Results:

• AEHMO demonstrated superior performance in high-dimensional optimization tasks on the CEC2017 benchmark suite.
• In a 150-node WSN scenario, AEHMO achieved an average energy consumption of 0.395 J, with First Node Dies (FND) at 1150 rounds.
• In a large-scale 1200-node heterogeneous WSN with a mobile sink, AEHMO significantly extended network lifetime, outperforming competing algorithms with an FND of 362 rounds and Last Node Dies (LND) at 56,946 rounds.

Conclusions:

• AEHMO offers superior energy efficiency, scalability, and adaptability for CH selection in both homogeneous and heterogeneous dynamic WSNs.
• The proposed algorithm effectively overcomes premature convergence and local optima issues prevalent in traditional metaheuristic approaches.
• AEHMO's integrated strategies enhance exploration, prevent clustering imbalances, and improve overall network longevity.