Genetic Algorithm with Heuristic Mutation for Wireless Sensor Network Optimization

Bio-inspired metaheuristics can be useful for the optimization of complex systems. Wireless sensor networks (WSNs) are massively distributed cyber-physical systems whose efficient operation requires appropriate design and control strategies. In certain contexts, like with randomly deployed WSNs, the physical network configuration can be affected only minimally, and optimal control strategies are crucial for optimizing network performance metrics like lifetime, coverage, and energy consumption. These metrics often conflict with each other, making network optimization a complex multi-objective problem. In this study, we introduce an improved version of a bi-objective genetic algorithm for the optimization of sensor network lifetime and target coverage. The new algorithm uses the generic evolutionary optimization framework together with a problem-specific heuristic mutation operator. We investigate the ability of the algorithm to find sensor schedules that extend network lifetime, and improve average target coverage while satisfying the minimum coverage requirement and show that the improved algorithm delivers better schedules than the original GA.