Abstract: Aiming at the problem of uneven energy consumption in wireless sensor network (WSN), on the basis of the uneven clustering, a kind of routing optimization algorithm for wireless sensor network based on fuzzy theory (RABF) is proposed in this paper. In the clustering stage, the relative surplus energy, the centripetal rate and the successful transmission rate are used as the election indexes of each cluster head. And the cluster head is selected by the fuzzy logic reasoning. In the inter-cluster routing stage, the indexes weights of cluster head node are determined through the entropy weight method, a fuzzy comprehensive evaluation method is used to obtain the forwarding probability of cluster head, the cluster head node with the largest probability is chosen as the forwarding node, and the date is then transmitted layer by layer through the cluster head node. The simulation shows that the proposed algorithm has better performance in reducing energy consumption and prolonging network lifetime.

Key words: wireless sensor networks, fuzzy logic reasoning, entropy weight method, forwarding probability, fuzzy comprehensive evaluation
无线传感器网络WSN(wireless sensor network)是由部署在一定检测区域内的具有感知、计算和通信能力的微型传感器节点所构成的无线网络,但WSN的传感器节点一般是由电量有限的电池供电,能量消耗直接影响传感器网络的使用寿命^[1],高效利用节点能量、延长网络生命周期依然是路由协议设计的重点工作。
 ,  ,  , 文献[2]从均衡节点通信负载角度根据主次参数的概念进行簇首选举,采用多跳方式与基站通信,克服了单跳通信方式能量消耗过快的不足,但多跳路由会使距基站近的簇首多次转发远离基站的簇首信息进而加剧节点死亡的问题不可忽视, 文献[3]与文献[4]通过调节簇半径的大小,采用非均匀分簇的方法减少靠近基站的簇的通信半径进而均衡网络负载, 文献[5]根据节点剩余能量对下一跳节点的选择方向动态调整,但这种算法易导致陷入局部最优,不能满足整个网络能耗的低能化要求, 文献[6]在簇首选举阶段考虑到相对剩余能量因子,控制低能量的节点成为簇首的概率, 文献[7]首先通过定义能量代价,选择能量代价最小的节点作为下一跳,但在区域约束问题上对于节点选择比较粗略,网络能耗较大, 文献[8]通过模糊理论与蚁群算法相结合的方法建立路由,该算法在网络节点的存活率方面有所提升,但权重赋值的主观性却使评判结果缺乏客观性。
 ,  ,  , 模糊综合评价法是用模糊数学对受到多种因素制约的事物或对象做出总体评价的方法,它将多个因素综合评价得出优先级,适合非确定性问题的解决。熵是对系统状态不确定性的一种度量,评价指标的熵越小,说明该指标提供的信息量就越大,在综合评价中的重要性就越强,权重就越高,反之越低。
 ,  ,  , 基于以上分析,在EEUC(energy-efficient uneven cluster )的基础上,本文提出一种基于模糊理论的无线传感器网络路由优化算法RABF(routing optimization algorithm for wireless sensor network based on fuzzy theory)。该算法采用模糊推理得出各节点的当选簇首的概率值,由熵权法确定簇首节点指标权重后,通过模糊综合评判法得出簇首的转发概率值,由概率值较大的簇首节点作为转发节点,以此延长网络寿命,提升网络性能。

1 模型假设