Bio-Inspired Energy-Efficient Routing for Wireless Sensor Networks Based on Honeybee Foraging Behavior and MDP-Driven...

Area of Science:

• Computer Science
• Artificial Intelligence
• Network Engineering

Background:

• Wireless Sensor Networks (WSNs) face challenges with limited node energy and variable data traffic.
• Existing solutions struggle to balance energy efficiency and network longevity in dynamic environments.

Purpose of the Study:

• To propose a bio-inspired hybrid optimization framework for mobile sink trajectory planning and adaptive routing in WSNs.
• To address energy constraints and spatiotemporal data heterogeneity by integrating MILP and MDP with Q-learning.

Main Methods:

• A hybrid framework combining Mixed-Integer Linear Programming (MILP) for global planning and Markov Decision Processes (MDP) with Q-learning for local adaptation.
• MILP determines optimal cluster head access sequence for the mobile sink, while Q-learning optimizes routing parameters in real-time.
• An epsilon-greedy strategy and periodic decision updates balance exploration, exploitation, and computational efficiency.

Main Results:

• The proposed MILP-MDP framework significantly improves network lifetime compared to baseline algorithms.
• Demonstrates superior energy balance across network nodes.
• Effectively manages mobile sink trajectory planning and adaptive routing in dynamic WSNs.

Conclusions:

• The integration of bio-inspired foraging strategies and reinforcement learning offers an efficient and robust solution for WSNs.
• The hybrid framework effectively tackles the dual challenges of energy constraints and data heterogeneity.
• This approach provides a promising direction for optimizing WSN performance in dynamic conditions.