EFFICIENT BROADCAST SCHEDULING BASED ON FUZZY CLUSTERING AND HOPFIELD NETWORK FOR AD HOC NETWORKS
Efficient broadcast scheduling in Ad hoc networks is important to avoid any conflict and to exploit channel resource efficiently.The broadcast scheduling problem (BSP) for Ad hoc is an NP-complete issue.In this paper, combination of fuzzy clustering and Hopfield neural network (FC-HNN) technique is adopted to solve the TDMA (time division multiple access) broadcast scheduling problem in Ad hoc.We formulate it as discrete energy minimization problem and map it into Hopfield neural network with the fuzzy c-means strategy to find the TDMA schedule for nodes in a communication network.Each time slot is regarded as a data sample and every node is taken as a cluster.Time slots are adequately distributed to the dedicated node while satisfying the constraints.The aim is to minimize the TDMA cycle length and maximize the node transmissions avoiding both primary and secondary conflicts.Simulation results show that the FC-HNN had superior ability to solve the broadcast scheduling problem for Ad hoc over other neural network methods as well as improves performance substantially in terms of both channel utilization and packet delay.
Tracking Broadcast scheduling Ad hoc network Fuzzy clustering Hopfield neural network
XI-ZHENG ZHANG
Department of Computer Science, Hunan Institute of Engineering, Xiangtan, 411104,China;College of Electrical Information Engineering, Hunan University, Changsha, 410082,China
国际会议
2007 International Conference on Machine Learning and Cybernetics(IEEE第六届机器学习与控制论国际会议)
香港
英文
3255-3260
2007-08-19（万方平台首次上网日期，不代表论文的发表时间）