传感器节点能量有限一直以来都是无线传感器网络的关键所在.针对该问题对传统的GAF(geographic adaptive fi-delity,GAF)算法进行了改进.改进的GAF算法引入了支持向量回归机(Support Vector Regression,SVR)来优化虚拟单元格的划分,同时将正方形网格改为圆形区域;另外,通过改变圆形区域的半径来加强相邻区域的连通性.结果显示,与传统的GAF算法相比,改进后的算法具有更大的优势,降低了节点能耗.

The uneven distribution of cluster-heads and the uncertainty of the number of cluster-heads in every loop cause defects in the LEACH algorithm.In this paper,the optimization from two different aspects based on the above problems is proposed.The first one is the optimization based on the SVM,that is,taking the location of the node into consideration and searching for supporting vectors to divide the wireless sensor network(WNS)into several zones,which can relieve the inefficiency caused by the unbalanced distribution of the cluster heads.The second one is to improve the cluster-heads selection mechanism in the LEACH algorithm,that is to say,solving inefficiency caused by the uncertainty of the number of cluster-heads in every loop through defining the number of cluster-heads in every single loop.From the analysis of the statistics and data collected from the simulation experiments of the improved LEACH algorithms,the conclusion can be drawn that the optimized algorithm can prolong the life cycle of WNS effectively as well as decrease the energy consumption of WNS obviously.