Studying Upper Bounds on Sensor Network Lifetime by Genetic Clustering

In this paper we propose a novel mathematical model for calculating the upper bounds on the lifetime of a sensor network. Sensors are organized into clusters and a linear programming model is introduced for calculating a cluster head rotation schedule. Unlike most other clustering algorithms, our algorithm maximizes the network lifetime rather than minimizing the energy dissipation of sensors. Compared with previous models, our approach can adapt to heterogeneous sensor networks with non-uniform transmission rates and battery power. In addition, the new protocol can maximize the K-of-N network lifetime, which has not been studied by most previous approaches so far. We use a genetic algorithm to compute optimal cluster formations. Simulation results show that our model can extend the lifetime of a sensor network up to five times over that of existing approaches.