基于蚁群算法和BP神经网络的信道分配策略的研究.pdfVIP

第29卷 第3期 传 感 技 术 学 报 V01．29 NO．3 2016年3月 CHINESEJOURNALOFSENSORSANDACTUAT0RS Mar．2016 ResearchonChannelAllocationStrategyBasedonAnt ColonyAlgorithm andBPNeuralNet．work一 ZHAIXueming，WANGJia，LIJinze (SchoolofControlandComputerEngineering，NorthChinaElectricPowerUniversity，BaodingHebei071003，China) Abstract：Themainpurposeofstudyingchannelallocationstrategyofwirelesssensornetworkistoimprovenet— workthroughputandcapacity，reducetransmissiondelayofthenetwork，andusethelimitednetworkbandwidthre— sourcesefficiently．Applicationofmulti—channelMAC protocolcansolvetheproblem ofnetworkparalysiscaused bychannelinterference．Duetothecharacteristicsofthewirelesssensornetwork，thisarticleproposedadynam— icfeedbackloadbalancingchannelallocationstrategybasedonantcolonyalgorithm．ThisstrategyappliedBPneural networktopredictthechannels’load，thenfilteredthechannelsbyusingtheload—balancingalgorithm basedonant colony，finallyallocatedthechannelsequencesbyusingthemaximum discretechannelallocationalgorithm．Final— ly，thisprotocolwassimulatedandimplementedonNS2platform，anditwasalsocomparedwiththeMMACprotocol andtheSMACprotoco1．Accordingtothesimulation，theMACprotocolweproposedinthispaperperformedbetter thantheMMACprotocolandtheSMACprotocolintermsofnetworkthroughput，networkdelaytoalargedegreeof improvement．Itcaneffectivelyreducethenetworkdelay，increasenetworkthroughputandenhancetheanti—jam— mingcapability． Keywords：wirelesssenornetwork；multi—channeldistributionmechanism；antcolonyalgorithm；BP neuralnet— work；maximum discrete；NS2simulation EEACC：1295；6150P；7230 doi：10．39690．issn．1004-1699．2016．03．024 基于蚁群算法和BP神经网络的信道分配策略的研究 翟学明 ，王 佳，李金泽 (华北电力大学(保定)控制与计算机工程学院，河北保定071003) 摘 要 ：研究无线传感器网络信道分配策略的主要 目标是提高网络吞吐量和容量，减小网络的传输时延，最大限度的利用有 限的网络带宽资源。多信道MAC协议的应用，可 以有效地提高网络通信的可靠