Optimized Back Propagation Neural Network Using Quasi-Oppositional Learning-Based African Vulture Optimization...

Area of Science:

• Computer Science
• Electrical Engineering
• Network Engineering

Background:

• Wireless Sensor Networks (WSNs) face energy efficiency challenges due to redundant data and inefficient routing.
• Limited battery capacity in sensors necessitates advanced optimization techniques for prolonged network operation.

Purpose of the Study:

• To propose an optimized data fusion and cluster-based routing mechanism for WSNs.
• To enhance energy efficiency and extend the operational life of WSNs.

Main Methods:

• Developed a Quasi-Oppositional Learning (QOL)-based African Vulture Optimization Algorithm (AVOA), termed QAVOA.
• Integrated QAVOA with a Back Propagation Neural Network (BPNN) for optimizing weights and thresholds to reduce data redundancy.
• Implemented QAVOA for optimal Cluster Head Node (CHN) selection and shortest path route discovery.

Main Results:

• The QAVOA-BPNN method effectively minimizes node energy consumption through data fusion and optimal routing.
• Achieved a significantly higher network life expectancy (4820 rounds for 500 nodes) compared to existing CL-HHO and ISSDE methods.
• Demonstrated improvements in throughput and Packet Delivery Ratio (PDR), while reducing End-to-End Delay (EED) and Packet Loss Ratio (PLR).

Conclusions:

• The proposed QAVOA-BPNN offers a superior approach for energy-efficient routing and data management in WSNs.
• This method provides a substantial increase in WSN longevity, addressing critical limitations of current technologies.
• QAVOA-BPNN is a promising solution for environmental monitoring and tracking applications requiring sustained network performance.