Abstract: This paper presents a 2-Level Flow Adjustment(2-LFA) mechanism using Autoregressive Integrated Moving Average(ARMA) flow forcasting model. Nodes can periodically forecast the next phase’s flow according to this model, and compare the value to the defined flow thresholds, and make the decision on how to transmit packages. The over-flow node avoids its fast death by establishing a temporary virtual circuit between low load brother node and high load son node. Simulation result verifies that 2-LFA can considerably prolong the lifetime of bottle-neck nodes and shorten the lifetime difference among the whole network, compares to 1-Level Flow Adjustment(2-LFA) and none adjustment mechanism, balances the energy consumption at the same time.

Key words: energy consumption balance, AutoRegressive integrated Moving Average(ARMA) model, 2-Level Flow Adjustment(2-LFA), bottle neck node, OMNeT++ simulation verification

摘要： 结合自回归滑动平均(ARMA)流量预测模型，提出一种二级流量调节机制(2-LFA)。节点根据模型周期性地预测下一阶段流量，与预定义的2级流量阈值作比较，选择执行最恰当的数据转发方式。过载节点通过在低负载兄弟节点和高流量子节点间建立临时虚电路，避免因流量过大而快速死亡。仿真实验结果表明，与一级流量调节和不带调节机制的路由协议相比，2-LFA可以有效地延长瓶颈节点的生存时间并缩小网络节点生存时间的差异，在延长网络寿命的同时均衡节点能耗。

关键词: 能耗均衡, 自回归滑动平均模型, 二级流量调节, 瓶颈节点, OMNeT++仿真验证

CLC Number: 

• TP393