資源描述:
《多源組播網(wǎng)絡(luò)編碼的優(yōu)化構(gòu)造-研究》由會(huì)員上傳分享,免費(fèi)在線閱讀,更多相關(guān)內(nèi)容在應(yīng)用文檔-天天文庫(kù)。
1、摘要隨著信息時(shí)代的到來(lái),人們對(duì)網(wǎng)絡(luò)通信質(zhì)量的要求越來(lái)越高。網(wǎng)絡(luò)編碼的提出是網(wǎng)絡(luò)通信研究領(lǐng)域中的一項(xiàng)重大突破,其中,面向組播通信的網(wǎng)絡(luò)編碼技術(shù)已成為網(wǎng)絡(luò)通信研究領(lǐng)域的重要課題之一。針對(duì)多源組播連接問(wèn)題,在假定無(wú)環(huán)的情況下,在基于單目標(biāo)優(yōu)化達(dá)到網(wǎng)絡(luò)最大吞吐率的網(wǎng)絡(luò)編碼構(gòu)造方法的基礎(chǔ)上,本文提出了使各子圖的組播容量組成的向量達(dá)到最優(yōu)的網(wǎng)絡(luò)編碼構(gòu)造方法。對(duì)多源組播網(wǎng)絡(luò)進(jìn)行子圖劃分,每個(gè)子圖對(duì)應(yīng)一個(gè)單源組播網(wǎng)絡(luò)。不同的子圖劃分方式能得到不同的單源組播網(wǎng)絡(luò)組,各組中的單源組播網(wǎng)絡(luò)的組播容量互相制約。為了求得最優(yōu)的子圖劃分方式,本文采用粒子群優(yōu)化算法對(duì)子圖進(jìn)行劃分,并動(dòng)態(tài)求解包含組播容
2、量向量的pareto解集。Pareto最優(yōu)解集中的pareto解對(duì)應(yīng)最優(yōu)的幾種子圖劃分方式。若用戶優(yōu)先考慮某個(gè)單源組播網(wǎng)絡(luò)的組播容量,可以在解集中選擇合適的解向量進(jìn)行線性網(wǎng)絡(luò)編碼構(gòu)造。分析表明,在網(wǎng)絡(luò)吞吐量方面,該方法優(yōu)于傳統(tǒng)路由傳輸技術(shù)。針對(duì)有向有環(huán)的多源組播網(wǎng)絡(luò),本文提出了一種網(wǎng)絡(luò)編碼的構(gòu)造方法。首先采用深度優(yōu)先遍歷方法對(duì)組播網(wǎng)絡(luò)進(jìn)行搜索,找出其中所有的環(huán),并求得各個(gè)環(huán)的長(zhǎng)度。然后依次在每個(gè)環(huán)中去掉一條有向邊,在網(wǎng)絡(luò)編碼構(gòu)造時(shí),這些邊不參與線性網(wǎng)絡(luò)編碼。這樣有向有環(huán)網(wǎng)絡(luò)被轉(zhuǎn)化成了有向無(wú)環(huán)網(wǎng)絡(luò)。不同的去邊情況能夠形成不同的有向無(wú)環(huán)網(wǎng)絡(luò)。最后求得各有向無(wú)環(huán)網(wǎng)絡(luò)的組播容量,
3、找出具有最大組播容量的無(wú)環(huán)組播網(wǎng)絡(luò)進(jìn)行線性網(wǎng)絡(luò)編碼構(gòu)造。仿真測(cè)試結(jié)果表明,提出的方法是可行的。關(guān)鍵詞網(wǎng)絡(luò)編碼,組播,有環(huán)網(wǎng)絡(luò),深度優(yōu)先搜索HABSTRACTWiththecomingofinformationage,peopledemandahigherqualityofnetworkcommunication.Networkcodingtechnologyisa.majorbreakthroughinthefieldofnetworkcommunication.Networkcodingbasedonmulticastcommunicationhasbecomeoneo
4、ftheimportantproblemsincommunicationresearcharea.Aimingatmulti—sourcemulticastconnection,thisthesisproposesanetworkcodingconstructionmethodtoachievethemaximumvectorscontainingmulticastcapacitiesofsub—graphs,assumingthatthereisnoring.Thismethodisbasesonsingle—objectiveoptimizationtoachieve
5、themaximumthroughputofthenetwork.Firstitdividesthenetworkintoseveralsub—graphs,whicharesingle—sourcemulticastnetworks.Therearedifferentgroupswhicharemadeofsingle—sourcemulticastnetworksindifferentdivisionmethods,andthemulticastcapacitiesconstraineachotherinthesamegroup.Inordertoobtaintheb
6、etterdivisionmethodtomaximizethemulticastcapacityofeachsingle-sourcemulticastnetwork,thispaperusesparticleswarmoptimizationalgorithmtodividethenetworkintosub—graphs,andsolvestheparetosolutionsetwhichcontainthemulticastcapacityofeachsingle—sourcemulticastnetworkdynamically.Theparetosolutio
7、nvectorsintheparetosolutionsetcorrespondwiththeseveraloptimalizingdivisionmethods.Ifuserstakeintoaccountthemulticastcapacityofcertainsingle—sourcemulticastnetworkfirstly,theycanchoosetherightsolutionvectorfromtheparetosolutionset,thenconstructthelinearnetworkcoding.