Tài liệu Bài toán định tuyến tối ưu trong mạng viễn thông Việt Nam. ppt

8 542 2
  • Loading ...
1/8 trang

Thông tin tài liệu

Ngày đăng: 27/02/2014, 06:20

Te.pchi TinliocviDietlkhienhoc,T.17,S.l(2001),46-53", ~,<A , ABAI TOAN DINH TUYEN Tal U'U TRaNG MANG VIEN THONG VIET NAM. . .DOTRUNG TA, LE VAN PHUNG, LEBACKIEN. Abstract.The purpose of this paper is to look at the optimal routing problem in Vietnam telecommunicationnetwork, in which we can approximately solve it, using the method of penalty and gradient functions.Torn uit.Trong bai nay chung toi decapden bai toan dinhtuy en t5i U"ll trong m~ng vi~n thongcrVi~tNarn. Chung toi girl.ixfip xl baitorinnay d1!-'atrenphuo'ngph ap ham ph at ket ho'p vO'i gradient.1. M()" DAUXfiydung va giai bai toan dirih tuydn toiU'Ula viec rat din thiet trong corigtac thiet ke va quihoach mang vi~n thong, nhfit la nhirng mang vi~n thong moi ph at tri~n nhu mangViet,Nam. Ngu'oithiet H, quan tri m<).ngphai xay dung ham muc tieu , phii ho'p vo'i d~c digm m<).ng,hru hrong clingnh ir rnuc dich toi tru dii-t ra. Vandedat ra la can 11!achon phuo'ng ph ap thich hop nhu: phtro'ngph apxfipxi lap dg giii bai totin nay. N9i dung bai viet nhlm dua ra mot each giii bai toan dinhtuyd n toiU'Umo hin h m<).ngvi~n thong Viet Nam nho phtron g phap ham ph at Ht ho'p v6i gradient.2.L1J"ACHQN MO HINH M~NG - LUU LUQNG vA XAY DVNGHAM MVCTUtUHien nay, trong cac cong trlnh ngh ien cuu va thuc te cac l11<).ngvien thong hien dai tren thegioi , mo hlnh mang khorig ph a.i cap thuo ng du'o c chu trong do c6 nhieuU'Udigm so voi m ang ph ancap. Tuy nhien , thuc te m<).ngvi~n thong Viet Nam cho th ay ring trong nhicu nam toiday mo hlnhmang ph an cap (it nhat lit hai cap) vin se Lontai. Xuat ph attir kh a niing trrig dung vaynghia t.hucte cila bai to an dinh tuyen toiU'U,chung ta IU'a chon mo hln h mangphfinhai cap: cap1bao gomm t6ng d ai lien tIn h va cap 2 bao gomnt6ng dai Host noi hat, tatdcac luu hrong lien tlnhxufitphattir cac capthfipho:n (cap 3) deu du'o'c coi laxufitph attir cac t6ng dai Host noi hat (hlnh1).ffJNi.J;BiJoHinh1.Mo hlnh dinhtuyeri trong mang hai capBAI ToAN DINH TUYEN TOI UU TRaNG MANG VIEN THONG47Cap 2 Ii 11O'ixu at ph at nhu diu luu IU'9ng vi ciing Ii noi. ket th uc cii a ltru IU'9'ng [dich den)' docacluu hro'ng deu co hurrngxuat ph attirnut di toi. nut dennent5ng corig ta co n(n- 1) nhu di.uluu IU9'ng Aii,tuo'n g irng voi n(n-1) c~p t5ng d ai di-deni-], (i= 1 ;- n,] = 1 ;- n,i=I- ]),Ta xemxetquatrtnhx11,11cac nhu cau hru hrong AiJ,Tru'o'chet hru IU'9ng Aii duocduavao tuyeri trungke noitruc tiep giiia hai nutivi]voidung hrong Nil.Cac luu luo ng Aii co ph an bo Poisson, Xac sufit cuoc goi bi chan tr en tuyen n aydu'oc xac dinhb~ng cong thuc Erlang-B n htr sau[1]:E(A,N)=ANIN!NL(AtIt!)i=()(1)Nlur vay, phlin luu hrong duo c luutho at se la, AiJ(1 - Bii) vi ph an hru hrorig bi rig hen lai IiAiJBiJ, Ta kihie u phan hru IU'9'ng bi ng hen n ay Ii aiJ; aiJ=AiJBii, Phfin hru hro ng aiJ dU'9'C goiIi hru hro'ng tr an vi d u'o'c phan chiath an h cac phfin n ho dg du'a len cac t5ng d ai lien t.inhkvoicacxac su at (ty Ie) a~J, Lka~=1.Cac phfinluuhro'ng n ho aiia~ diro'c dua len c ac tuyeri trung keNuik, dU'9'C chuye n m ach qua c.ic t5ng d ai lien dnhkv a du'a xuong cactuye n trung ke N dkidg toicac nut dich den ], Ta kih ieu Buikla xac su at bi nglien rn ach tren tuyen trung ke Nuik, con Bdkila x.ic suat bi ng hcn m ach tren tuydn trung ke N dki, Xac suat cuoc goi dtroc ket noi qua t5ng d ailien tinhkse Ii xac su atd. hai deantuydni-kvik-]deu co it n hfit mdt ken h con roi vi b~ng(1- Butk)(1_ Bdki), Nh ir vay phfin hru hro ng bi tr an co huangi > ] di qua t5ng dai lientinhktoiduo'c nut dich ] se la:(2)Moi mot donvihru hrorig co hu'ongi > ] khi diroc ket noi qua t5ng dai lien t.inhkneu denduoc dich ] se mang lai mot loi. Ich la w~i, w~i c6 th€ Ii nhu nh au cho moik(vi du khi w~J chinh la.currc phi lien t.in h] ho~c kh ac nhau thee k (neu la w~i lo'iIch rong sau khi lay doanh thu tr ir di cacchi phi lien quan), hoac w~J c6 thg la trorig so cil a du'ong thOng doivoihru IU'9ngi > ] do ngu'oiquan tri m;;tng dat ra nhiim rnuc dfch d ie u khi€n luu hro'ng , 6' dayta lay t.ru'ong h9'P chung nhatkhi w~i la trorig so - 100iich cu a duo ng thong, Muc t.ieu cu a bai to an dat ra la xac dinh cac ty I~a~sac cho t5ng 19i. ich mang lai tren toan m angtir cac hru IU9'ng den duoc dich(g~)Ii IO'n nhfit [2],tiic la:m ax L L Lw~g~hay Iiikmax L L L w~JaiJa~i(1 - Buik)(1 - Bdki),ik(3)Day chinh la ham rnuctieu mata can toi uu ho a thee c ac biena~,Ta xac d inh c ac ring buoccu a ham muctieu nay, Nh u dil neucrph an tr en, c ac du'ong thong th u cap qua t5ng d ai lien t.inhkbao gom h ai doantuydn i=k: vak-],Tuy hai doan tuydri nay Ii di?c lap v oi nhau n hung xac suatcuoc goi bi ch iin tren cac dean tuyen nay (Buikv a Bdki) lai lien quan ch~t che voinhau vi cu oc goichi c6 thg ket noi du'qc khid,hai doan tuyen deu c6 it n hfit mot duo-ng thong can roi, Ne u xet trenbuc tran h t5ng th€ luu luong vi mang thi hru hro ng di tren bat cii doantuyen s n ao deu ph ai Iit5ng cu a tat cac ph an hru hrong c6 11U'6'ngi-]kh ac nhau nhung cling chung dean tuydn s do, Ketqua Ii xac sufit ng lien m achB"tren doan tuye n s do cii ng phu thuoc v ao xac suitt ng hen m ach cti acac dean tuyen kh ac trong cling m a tri).n dirong thongX,dm a c ac hru hrongi-]do qua, Trong[1],Girard dil chung minh r5.ng, trong mo hinh m~ng heat dong theo nguyen Iy chia t3.i, quan h~ giii'acac xac suat nay dU'9'C bi€u di~n b oi h~ phuong trlnh "di€m bat di?ng Erlang" - Erlang Fixed-PointEquation:(4)Ma tran thongX",IIi ma tran gom cac phan tu'X.,1bhg 1 ho~c 0, bi€u thi r5.ng do~n tuyensc6 n5.m trong du'ong thongIhay khong, Can Alia luu luong dau vaG crta du'ang thongl.Trong48DO TRUNG TA, LE VAN PHUNG, LE DAC KIENtru'ong hop cua .ta, tatd.cac du'o ng thong i-k-J bao gom hai doan tuyen, rien h~ phtro'ngtrrnhdifm bat dong Erlang se chi bao gom hai bigu thirc lien quan den Buikvaiu»,Ta b5 sung themcac rang bU9C co lien quan t&i cac h~ so chia til.i a~J va ham Erlang-B, va viet lai ham m~c tieu nhirsau:maxF=LI::aiJ{Lw~J(1 - Buik)(1_ BdkJ)a;}iJkvoi cac rang bU9C .Bik = E(L[aiJa;(1 - BdkJ)J,Nuik),JBkJ =E(L[aiJa;(1 - Buik)J, NdkJ),,m'\' ii c:ak=1, a'J > 0k=lk - ,E(A,N)= ANIN!N.2:(AtIt!),=0(5)Vi=1 -:-n,VJ=1-:-n,i-=IJ;Vk=1 -:-mva cac tham so dau VaG saum - so hrong cac t5ng dai lien tlnh cap 1,n -so hro ng cac t5ng d ai n9i hat cap 2,AiJ - nhu cau hru hro'ng xu at ph attir nutidg di toi nutJ',NiJ - dung hro ngtuydn trung ke noi tru:c tiep hai nutivai,N uik- dung hrong tuyen trung ke i-ktir nut diilen t5ng dai lien tln h k,N dkJ- dung hro'ng t.uye n trung ke k-Jtir t6ng dai lien tlnh k xuo ng nut deni,w; -loi ich mang lai t.ir mdt don vi hru hrong huo'ngi >Jdi toi diroc nut denJb~ng dtrongthOng i-k-j di qua tong dai lien tlnh k.Tir ket qua giai bai toan toi uu(5),ta se co dtro'c mot b9 h~ sophfinchiaa;toi U'Ude' ap dungcho cac nut tong d ai diivoimuc dich mang lai IQi ich Ion nhat tr en to an m ang .3. SlJ-DVNG PHU'O'NG PHAp HAM PH.~T KET HQ'p GRANDIENTHEGIAI BA! ToAN QUI H04-CH PHI TUYEN(5)De' gitti bai toan(5),ta du'a ve bai to an qui hoach phituyeri (QHPT) d ang tong quat:F(X) >minH(X)=0 (6)C(X)::>0vo'i X=(a;,e.»,BdkJ) la vecto: trong khc ng gian m.n. (n +1)chie u.Crday t.a coi lucri Buik, BdkJla cac bien cantlrn .Ham so Erlang-B diidu'ccchirng minh la ham loi nhung di'eu kien nay chira dude' du'a bai toan(5)ve bai toan qui hoach loi. Tuy nh ien , cluingtathfiyham muc tieu va cac hamA .~ ,bAI' khl . , . d h'3F(X) 3H(X) 3C(X) • 1 (1)tren mien rang uoc aa.VI v a cac ao am ; ; V01J= , ,m.n. n +. . 3~ 3~ 3~hoan to an xac dinh du'oc. Ben canh do rang buoc bat d1ng th irc chi gom m.n.(n -1)d ang don giana;::>0la m9t IQi the trong qua trinh gi3,i khi tuyentinh hoatai m.3i bu'oc l)!.pr[ph uong ph ap xapxi QHTT (qui ho ach t.uydn tinh )) hoac khi xay dung cac ham ph at va tinh gradient [phtrong phapham ph at ket hQ'P gradient).BAr TOA.N DINH TUYEN Tor ULJ TRONG MANG vrEN THONG49Ta so sanh vi IU'achon phu'ong ph ap toi U'Ude' giai bii to an QHPT (5). Cac ph iro'ng ph ap QHPTthirong dung nhfit Ii phu'o'ng ph ap nh an tti: Lagrange, cac phu'o'ng ph ap hu'o'ng co the'[phuong ph aplnrong chap nh an duo c va phuong ph ap Frank- Wolfe)' phuo-ng ph ap Monte-Carlo, cac phiro ng ph apxfip xi (xap xi QHTT) va phuo'ng ph ap ham ph at , hoac ham ph at ket h91> voi gradient [3,4,51, Tuynhien, 2 pluro ng ph ap dau tien khong hie u qui eho bai toan khong IOi, phirong ph ap Monte-CarloSo' dothuatgiaibai toanQHPTTim ['={pIf1' :::::O}~ l" ~(pIf"1,)<0)Tinhcacham phat:F(z)->min,zEORN{f1'(z)<0,p =1,Pr' (z)=0,q=1,Q[Xay dirng{zv}->z*-opt)1P(z)=F(z)+-P'(z)+e" P"(z)e'P' -ham ph at ngoaiP" - him ph at trongLiiptiepe''=11\7FII11\7pI/IIe;v+l:=Qe;"e':=a'e'e":==a/lell5Kie'm traCtJ ::; C"'(hillll()?Z1):==zv:=v+1Giambu-octvtChonZOEORN.0. 0.' 0."EO (0°5).)"),)f3EO(0,5,0,8),e;oEO (0,1,1)P'=L(r"(z))2+L[JI'(z)]2'-__ P_"_=_~__f~_,(~.•)__I'EI'II,EI" ~Kie'mtrav=O?5h (z)= ~[\7F (z)+ ~\7P' (z)+ e;"\7P" (z) ]s1/::,.=p(z+Ah(z)) ~ P(z)+'211hl12s50UOTRUNG TA, LEvAN PHUNG, LEUACKIENrat thich hop voi cac bai to an QHTT vo'i ham m~ctieu va cac rang bU9C khorig loi cling khorig lorn,nhtrng bi gi6'i han bo'i so bien'S 30, Ph uong ph ap xap xi QHTT cho chungta gi3.i l~p bai toanQHPT mot each dongian hon , tuy n hien viec thu'o'ng xuyen ph ai ki~m tr a tinh chap nh an du'occu a di~m xufit ph attai m6i biro'c l~p se lam cho bai to an tr6' rien cong kenh va giarn toc d9 h9itu ,Trong phtrc ng ph ap ham ph at ket hop vo'i gradient viec dung gradient lam huong di toi iru trongm6i buo'c la.p se lam tang dang ke' toc do "tut' cu a gia tri ham m~ctieu, ben canh do cac ham ph atse lam cho mien xet ngh iern co hep t6'i rmrc co th~ v a luon luro'ng v ao trong mien chap nhan dtro'ccu a baitorin .6day do phuo'ng an xufit ph at khorig bi rang buoc ph ai th uoc mien chap nh an , segiarn nhe du'o'c nhie u ph ep ki~m tra nen toc d9 h9i tu tang nhanh dang k~ so vo'i phuo ng ph ap xapxi QHTT,Phuong ph ap ket ho p gradient va ham ph at la mot ky thu~t t&ng ho'p, ph at huy dU'C?,Cthe rnanhve toc do hoi tu nhanh cu a phiro ng ph ap gradient, loai bo diroc cac rang buoc phtrctap riho cacham phat d€ gi3.i baitoan QHPT dang t&ng quat, Ben canh do, qua ph an tIch dang rang bU9C, thayr5.ng viec tinh toan cac ham ph at co th~ duo'c do n gian di rat nhieu v a thuan loi cho v iec gi3.i tr enmay tinh, chung toi hra chon phu'o'ng ph ap ket hop ham phat va gradient d~ gi<ii b ai toan (5),T'huattoan ket h9'P ham ph atva gradient diroc van dungnhir sau15]:Xet bai t.otin:millF(z)= -Lw~JaiJa;:(1 -Buik)(1 - BdkJ)i.J.kvoicac rang buocf1'(z) :_a~J'S0 v a'\' iJW' .c:ak-1 = 0,vt ,J = 1,nkBuik-E(2:aiJa;:(l - BdkJ), Nuik)= 0JBdkJ-E(2:aiJa;:(I- Buik), NdkJ)= 0tT'}(z) :Vk = 1,mVi,]=l,n,i=/J(7)(p=I,P, q=I,Q)bienz=(a~, Buik, BdkJ),+Buo:c0: ChonZOERN;a,a',a"E (0,0,5);f3E (0,5,0,8) voiN=m,n,(n+I),+Buoc1: Batz =zO,chonvongl~p v =0,+Buo:c2: Xac din h c ac t%p chi sol'={pE{1,2, "P} 1f1'(z)2':oj,I" ={pE {I, 2, "P} 1f1'(z)<o}.+Buoc3: Xac dinh cac ham ph at ngoai va trongHam ph at ngo aiQP'(z)=L(T'}(z))2+L(f1'(z))2(/=1pEl'=L (La;: -If+L[Buik-E(L aiJa;:(I- BdkJ), Nuik)riJk i.kj+L[BdkJ-E(LaiJa;:(I- Buik),NdkJ)f+L(a~J)2k,Jii,J,kirng voia;:'S 0,Ham phat trong:(8)BAr TO.AN D)NH TUYEN Tor UU TRONG MANG vrEN THONG51P"(z)= ~ __1_= ~ ~L.JI'(z)L.'JEI" k akl' ,1,1.irng vci a~>O.(9)+BU'6'c4:Neuv= 0 chuydri den buoc 5, neu kh ac t6"i biro'c 8.+BU'6'c5: TinhVF(z),VP'(z),VpIItZ).+BU'6'c6: Tinh e' =IIV P'(z)11e'' =IIV F(z)11IIV F(z)11 ' IIV P"(z)11 .+BU'6'c7: Chon co E (0,1,1).+Bu'6-c8: Tin hh(z)=-[VF(z)+f,VP'(z) +c"VP"(z)].+BU'6'c9: NeuIlh(z) II>e;chuyfin t6"i10,neukhac: ki~mtra di"eukiene., ::;s",neu dung thi STOP, Ht thuc thuat to an,neu khac , d~t:1':11+1=aCt!,1':'=a'e",e"=a"c"va quay lai buoc 2.Z1)==ZIv=v+1DatA=1.Tinh!::::.=p(z+Ah(z)) - P(z)+~llh(z)112voiP(z)=F(z)+~P'(z)+c"P"(z).2c'Neu!::::. ::;0 dat z = z +Ah(z)va chuye n to-i buo'c 8, neu kh ac d~tA={3Ava chuydnto-i buoc11.Th ufit toan se Ht thuc khi:e.; ::; e"vo'i c· du nho, chon truo c,ta chonZu= z lam phtrong anxfip xito-iU'U(dien ra khi thuc hien buo c9)va cho day{zu}h9i tu aenz"-opt.+BU'6'c 10:4.CHUO'NG TRINH MAy TiNHVAKETQUATHUC NGHIEM..Tren thuc te, bai to.in dinhtuyen toi uu co qui mo rat Ion vi ph ai giai bai tcan tren qui moto an mango Gi<l.su:m<.tng co qui mon=100,m=5thl chungta se c6 tatdn.(n-1).m+2.m.n=100. (100-1).5+2.5.100=50.500bien. So rang bU9C dang dhgtlurcIan.(n-1)+2.m.n=2.5.100+100 . (100-1)=10.900con so rang buoc dang bat dhgthirclam.n=5.100=500.V6"i qui mo bai tori.n Ian nhu vay can c6 nhiing chucng trinh tinhtoan tr en cac ph tro'n g ti~n hiendai nh u cac d an may tinh 16"n (microcomputer tr6' len}, dieu ma trong ph am vi de tai nghien ciiunay kh6 th uc hi~n duoc. Tuy nhien, bai toan co th~ diro'c mo phorig d~ gi<l.itr en may tinh PC voiqui mo nho ho'n m a van giii:du'ocynghia thuc te, xuat ph at t.ir rriuc tieu neu tren, chung toi chonqui mo mang cel"trung blnh nho voi: m = 7,n= 3.Ch uon g trinh may tinh giai bai toan QHPT duo'c viet b~ng ngon ngir b~c cao Turbo Pascal 7.0va chay tren PC. Chuang trtnh bao gom cac mo dun chinh nhirsau:- Mo dun nhap cac so lieu dau van duoidang tep *.txt.- Mo dun thutuc (procedure) tinh ham Erlang-B.- Mo dun tinh cac tap chi sol'va L" va tinh cac ham P' v a P",- Mo dun tinh gradientVF,VP',VP",- Cluro'ng trrnh chinh.- Mo dun Ht xu at dau raa~va tinh gia tri ham m uc tieuF(z-opt)duoidang t~p *.txt.Ketquat.hircrighiernmo pbong1.vesu:h9i tu cu a thu~t toan: toc d9 hoi tu cu a thuat toan phuthuoc vannhieu yeu to, trongd6 cac yeu toCO"ban nhfit la d9 xap xlc.yeu cau vaS1).·phiretap cu a cac rang bU9C keo theo viec52DO TRUNG TA, LE V AN PHUNG, LE DAC KIENtInh to an cac ham phat, trong d6 cac tham s5 deu ph ai l<lYra tv: cac mang (array), M9t kh5i hro'ngtfnh toan dang k~ nii'a cling du'oc dan h cho viec t.inh gradient tai m6i buoc lap, Do su: dung cacham phat , viec chon phu'o'ng an xufit ph at Zo khong bi rang buoc, tuy nhien ta c6 the' rut ngiin motso bu'oc lap ban dau blng each dua ra mot ph iro'ng an xufit ph at [chu yeu la b9a;:lnlm trongmien chap nhfin duoc. ChU"011gtr in hdiro'cch ay tren maytinh ca nhan PC cau hinh:bo vi xu.' Iftrung tam CPU Intel Pentium II -266MHz, dung IUQ'ngb9 nho32MB RAM voi th oi giantin h to ankhoang5gifiy.2.vehieu qua cii a dinhtuyen toi U'Utheo IQ'i ich: muctieu bai toan din htuyen toi U'Utrongtru'ong ho'p cu a chung ta la ph an chia cac nhu cau hru hro'ng lien tjnh xufit ph at tir 7 to'ng dai noihat tr an len 3 to'ng d ai lientinh mot each toi uu nhlm dat du'oc hieu qu a (19'i feh) cao nhfit Bencan h d6, 101 ich cu a ng tro'i su dung cling du'oc dam bdo.Bdng1. So sanh h ieu qua 2phtrong ph ap toi U'ULuu hrcng taiDinh tuyeri toi U'Utheo 191ich Dinhtuyen toiiru:Loss->minI%F"GaS"r,c-s,60%3.508,806 5,58E-43.466,5023,16E-1670% 4.080,9218,83E-44.042,2716,72E-12I I I I I I80%4.646,589 9,20E-4 4.605,810 1,760E-890% 5.171,4680,00175.145,6691,71OE-5100% 5.671,985 0,00295.649,277 0,0027110% 5.994,006 0,0385.980,846 0,034120% 6.120,640 0,0986.110,1280,093130%I6.175,0670,1616.163,599 0,155i.140%I6.205,2080,2176.191,082 0,2116,500,005.500,006,000,005,000,004.500,004.000,003.500,0060r.70%80%90%100'/0 lWOt120%130% ).40%Hinh.2. Thay do'i gii tri ham muc t.ieu theo rmrc do tai(%)Trong bang1ta thay ring khi nhu cau hru hro ng thap (tir90%tro:xudng}, vi du nhirv ao cacthoi gian khong cao die'm, thuat toin dinh tuydn t5i U'Uthong th uo ng theo hru IUQ'ng se san cacBAI TOAN DINH TUYEN TOI UB TRaNG MANG VIEN THONG53phan lu'u hro'ng vao cac duo ng thongMd at dtroc tl I~ t5n hao thap nhat; do v ay neu so sanh thlGoS">GoS/" Viec nay chtmg to dingth uat toan dinh tuyeri toi U'U theo 10'i ich da ph an chia toi daluu lu'ong v ao cac d u'o ng thong mang lai h ieu qua(w~J)cao, d&n to'i gia triham muctieu Ion 11O'nF">Fb,v a chap u h Sn xac sufit cuoc goi bi roi 16'n hon , nhungta clingthfiyGoS" v&n n ho ho'n tiI~ cho ph ep la0,01v a 101 ich cii a ng u'o'i su dung vin duoc dam bao.Khi nhu cau luu luo'ng tang len , m~ng tro nen qua tai, an h htrong qua lai lin n hau cu a c acdong hru hrorig tro' rien ro r~tthi s\l' k h ac biet, giu'a GoS" v a GOSbkhong con 16'n niia, trong khi doviec dinh tuye n toi U'U theo gia tri van chota gia tri ham m uctieu F" Ion ho n mac du de? chenh l~chciing giam di. Dieu nay ch tin g to rbg t5ng cong hru hrorig diroc hrutho at trong d. hai truo'ng ho-pla g'an n h u' nhau va tien dan den gio'i h an cho qua (through-put) cd a m<).ng hro i, n hirng dinh t.uyeritheo gia tri dii ph an chi a toi U'U cac nhu cau hru hrong vao c ac du:o'ng thong mang lai 191 ich cao hon.5,KETLUA NLan dautien , viecxfiyd u'ng va giai bai t.oan dinh tuy en toi U'U cho m<).ng vi~n thong Vi~t N amduo'c d~t ra v a giai quyet mot each doc lap, dua tren nhirng ket qua ngh ien cU'U va nh irng van decon rno tai thai di~m hien nay tr en the gio'i ve d inh t.uyen. Vi~c xay dung ham m uctieu va c ac rangbuoc phu hop voimo hlnh rn ang Viet N am trong ttrong lai gan, sau do viec lu a chon phu'o'ng ph aptoi U'U hieu qua dg giiith an h cong bai toan la n himg ket qua mang ca tinh ly th uyet va thuc ti~n,Qua t.huc nghiern mo phong , dinh tuydn toi Ull theo 101 Ich da chung to dU'9'C su: iru viet ve hieuqua so voi mo hlnh dinh tuyen toi U'U khong d u'a tren yeu to kinh te, Bbg viec hiro ng gia tr~ hamITI\lCtieu theo 100iich +max, dinhtuye n toi U'U theo 191 ich da mang lai hieu qua [Ioi Ich] cao nhatcho nh a khai th ac mang, dong thai van dam bao chat hro ng dich V\l chap nh an dU'9'C cho kh achhang, Dieu nay n oi len d,ng viec lu'a chon dinh tuydn toi iru theo loiich la huang di dung dh, phiih91> voi xu the ph at trign cong nghe v a dich vu h ien nay, khi 101 Ich cua nh a khai th ac khOng thgtach roi khoi lo'i ich cu a ng u'oisU' dung,Phuong ph ap ham ph at ket h01> gradient Ii phuo'ng ph ap hieu qua d~ giai bai toan d inhtuyeritoi U'U dang QHPT ph iretap b5.ng 19i. the ve toc de? "tut" nhanh cu a gradient va lo ai bo duoc cacrang buoc ph ire tap nho:c ac ham phatDac biet trong cac mo hinh m<).ng trung ke voicac du'ongthong bao gom 2 deantuy en , viec xay dung cac ham phat va tinh gradient duo-c don gih v a rutgcn dang kg v a se giam thai gian xu' Ii tinh to an tr en may tinh. Viecxfiyd\lng v a ch aythanh congchuo'ng trinh maytin h da chting to su h ieu qua do, Day cling la dieu kien th uan 19i. cho viec xaydung bai toan cho c ac mo hlrih m ang v a hru ltro'ng kh ac.TAlLIEUTHAM KHAO[1]Andre Girard, Touting and Dimensioning in Circuit-Swithched Networks, INRS Telecommuni-cations (1990),[2]Adre Girard, Revenue Optimization of Telecommunication Networks, IEEE Transactions onCommunications41(4) (1993),[3] Blii Minh Tri v a Btii The Tam, Cuio trinh TaiU'Uh.oa;NXB Giao thong V~n d,i, 1997,[4]Dimitri P. Bertsekas, Constrained Optimization and Lagrange Mutiplier Methods, Academicpress New York - London - Pari - Tokyo, ban dich tieng N ga, Nh a xu at ban Ph at thanh va Lienlac, 1987,[5]E, Polak, Computational Methods in Optimization, Mathematics in Science and Engineenng,Academic press New York - London, 1971.[6]Le Dlic Kien , Dinh tuyen toi U'U trong mang vien thOng, Hoi nghi Va tuyen ai4n tJ: to an. quacIan thu; VII, 1998,Nhiin. bai ngay S - S - 2000Vien Con.q nghe thong tm . du'o c chu trong do c6 nhieuU'Udigm so voi m ang ph ancap. Tuy nhien , thuc te m<).ngvi~n thong Viet Nam cho th ay ring trong nhicu nam toiday. acac dean tuyen kh ac trong cling m a tri).n dirong thongX,dm a c ac hru hrongi-]do qua, Trong [1],Girard dil chung minh r5.ng, trong mo hinh m~ng heat
- Xem thêm -

Xem thêm: Tài liệu Bài toán định tuyến tối ưu trong mạng viễn thông Việt Nam. ppt, Tài liệu Bài toán định tuyến tối ưu trong mạng viễn thông Việt Nam. ppt, Tài liệu Bài toán định tuyến tối ưu trong mạng viễn thông Việt Nam. ppt

Từ khóa liên quan

Gợi ý tài liệu liên quan cho bạn