退避算法在Ad hoc网络中的吞吐量性能评估*

针对现有的主要退避算法，通过一维马尔可夫链分析了指数退避算法、线性退避算法和多项式退避算法以及各自对应的有最大重传次数限制的退避算法在饱和Ad hoc网络中的吞吐量性能，为无线Ad hoc网络的MAC机制的性能优化提供了理论依据。分析表明，对重传次数进行限制会减小吞吐量，同时导致丢包率的增加。三种退避算法中，多项式退避算法的吞吐量最大，而冲突概率最小，指数退避算法的吞吐量和冲突概率在三种退避算法中均居于中间；此外，吞吐量随着竞争节点的个数增加而减小，其中线性退避算法减小的速度最快，指数退避算法次之。

Throughput analysis of backoff function in Ad hoc network

The collision probabilities of nodes in Ad hoc network influence the performance of the MAC protocol. Because the node in network accesses the channel randomly, the backoff functions were adopted to resolve the collision traditionally. This paper used a one-dimensional Markov chain model to analyze the saturation throughput in Ad hoc network for generalized exponential, linear and polynomial backoff functions and these backoff functions with maximum retry limit. The analysis provides a basis for the performance optimization of MAC protocol in wireless Ad hoc network.