Evolutionary computation for lifetime maximization of wireless sensor networks in complex 3D environments

Scheduling the operating mode of nodes is an effective way to maximize the lifetime of wireless sensor networks (WSN). For a WSN with randomly and densely deployed sensors, we could maximize the lifetime of WSN through finding the maximum number of disjoint complete cover sets. Most of the related work focuses on 2D ideal plane. However, deploying sensors on the 3D surface is more practical in real world scenarios. We propose a novel genetic algorithm with redundant sensor auto-adjustment, termed RSAGA. In order to adapt the original GA into this application, we employ some effective mechanisms along with the basic crossover, mutation, and selection operation. The proposed operator of redundant sensor auto-adjustment schedules the redundant sensors in complete cover sets into incomplete cover sets so as to improve the coverage of the latters. A rearrangement operation specially designed for the critical sensors is embedded in the mutation operator to fine-tune the node arrangement of critical fields. Moreover, we modify the traditional cost function by increasing the penalty of incomplete cover sets for improving the convergence rate of finding feasible solutions. Simulation has been conducted to evaluate the performance of RSAGA. The experimental results show that the proposed RSAGA possesses very promising performance in terms of solution quality and robustness.