Wireless sensor network reliability evaluation based on genetic algorithm and coarsening combination evidence reasoning

As for the extremely unpredictable factors of Wireless Sensor Networks (WSN) with constrained resources operating in an unattended mode in uncertain dynamic environments. The distribution optimization and behavior evaluation of sensor-network nodes is vital to reduce the energy consumption and ensure effective information acquisition in distributions sensor network (DSN). Because of Bayesian probability incapability of capturing epistemic uncertainty, one evaluation scheme based on Genetic Algorithm and Dempster Evidence Theory (GDT) is proposed to optimize the node distribution and coarsen, refine combining algorithm. GDT offers an efficient the optimization of node distribution and framework for uncertainty quantifying and partial knowledge processing with sharply decreasing computation complexity. Simulation result can optimize the node distribution in target area and reduce the network energy consumption and increase the whole coverage rate at a relatively low cost , this ubiquitous low-cost computation appropriate to flexible Wireless Sensor Network Management Protocol for prolonging live-time of sensor network.