The energy consumption of wireless sensor networks (WSNs) includes the hardware consumption caused by the normal standby of the devices such as sensors, microprocessors, and wireless transceivers and the software consumption caused by the software work of the protocol. In this thesis, we investigated a new and practical model to achieve the lifetime extension goal of the wireless sensor network through ON/IDLE scheme which combined with base station scheme. By achieving the optimal lifetime of the sensor networks, the optimal base station scheme will be the major influential factor. We refined the status of the sensors from unitary to binary which includes ON and IDILE two states. The investigation of these two states represents from the single sensor node perspective to the whole network perspective. We detailed a novel perspective to define the TiON as the total active time period of the individual sensors and TiIDLE as the total sleep time period of the individual sensor. Base on the conditional above, we formulated a max-minimum problem. By means of maximizing the lifetime of the least alive sensors then achieving the goal of the whole-lifetime extension goal. Through the extensive derivation and calculations, we simplified the formulations from a mixed integer nonlinear programming problem (MINLP) to the NP-hard linear programming problem. By achieving the optimized lifetime of the sensor network, the optimal base station scheme will be the major influential factor. In a real-world environment, a WSNs network SN device is usually not evenly distributed. Some areas will have the more intensive equipment, while other regional equipment is relatively sparse. Then we develop a heuristic algorithm called Clustering Divide and Brute Force (CDBF) comparing with the Divide and Brute Force (DBF) to enhance the performance of feasible solutions to the BSP problem, and analyze its complexity, in the subnet division according to its actual structure will be similar to the SN device in a subnet. By prior clustering, the stochastic subnet can be used to divide the unreasonable discrete points, which optimizes the performance of the DBF algorithm in the non-uniformly distributed circumstances.
展开▼
