Title:
URL Source: blob://pdf/a2e2f555-9013-43e9-a337-ec67ad1ca8fb
Markdown Content:
## Fag Inene Sqstms (FIs) Kuk- based systu.dusy moleb, and fusy erpeat ysims tg indeena Systemy (FIs). The key wit os a du3sy logc Sln is FIS. FIS "IF.... THEN' Kues along wcth connec oay K on AND. fos making necessay deasion zuls. The inut to FIs may be dussg o CAisp , but te atput taom FIs in aluays set. gnnally known as Constuc ton 4wonking paincipk a< FIS. |Fgitcahon inpuat In<eene Caisp)it DataRule 4|base base knowledge base Decision - Making unt aDefiggkatonatpd Tnje aence (Cp) Fuzag Block diagram Qf FIS. |unit A AS constucttd ) ve dunctiol bleeus )A Aue bae tht costains IF: THEN )A data base that dekines the Q4 d34 set wsed in dussy 3) 9) Desision-making unt that pejoum opescaton the Ruley. memben ship tnchons kus. 13 9 Fugijicahon inexenca unit that Conveat the caus quanities into dugs4 quastitie in 5) De tggification iexena unit ttt conv ith ths figgy quontitiu into cisp quanti tiey Iritially, in the duzaiticaton untt, the caisp inpt is convea td into duzqH inpct. Vaiow tusshcato nethods aue emploed o this. A4tex this paoess; &ule based u douad Bat base and Kule bue au collec tively called kaowledqe base Fndlly. de<uzatication paoes1 i (amied out to produc Cisp adput. Mainly the dusy uls cau lonud to the aule base and Suitable deas ion aae made tn the decisin - making nit. nlhey au: tht contains numekoLus d434 Ruls. Methods Q FIs Tvu cu two important types af FIS I Mamdani FIS Suqono FS. lhe dyeAeNCe by the tuo method lis is the Consequnt of fugay Fuz3y sets ae wed as kule (onseqest in Mamdani FIS ule - Linzan guncton o nput vcaiables ae ucd ag Rule consequnts is Sugeno's maihod. Mamdois kule dind a qeatex acceptante in al niversai oppzozimatoxs than Sugeno's mocal Mandani FIS. p menbership funchions au ezpected to be fay sets. Ajtex aggagaton paoas, each cutpud vauable conitains a duaH ets, hene deyugifcaton is cmpoxtant at the otpd stge steps to comput otpud jom this FIS. kp-I: Deleamine a Set a< tg3y kules. ep-: Nake the ifp tgg4 using tp merdbership fuxhony step 3 : Combine the duajied pdg kues oa establi shing a ul stengy stap4: Dete Amin tha consequant q the ule by 1sup 6 :Finally obtoinec. acconding comlbining tha Rule staength ond the output monsbekship duncton. stap5: Conbine all the Consequnta to get as ofp distaibution. 3Fpd distaibution Ca de<ajed ocdpt distai bution 1Rule sBungih Jand to the I Momdani FIS.I Rule Then stugth ouput clistaibutn 1Then Consider a too-inpud Madanu FIS wth tu00 ules The modal dusifieu the hwo inputa by inding tte inteasec tion a wo (isp ingut valua. wcth input membeaship unctin . The minumm onoaation is wed to Computu the dussy input ind or conmbining he too daated input to obBas kue stengh. The output mambeaship jncton is cipped at te suk stength. Fioaly operaba is wed to comput the #u marimun the two ules >The duz34 kue aue state ment and AND /OR COnnectves .( The consegueny the ule can be obtaind in too steps: joumed wing E-THEN' ) by computig the kule sthengih completely wsing the tusziyed yp faom the fuzy Combitoy a) bg cippcg the ouput membexslaip unchon at the aul The outputs a al the jusgy Reule Rles aa (ombid to cotais ona jzg otput distibuton Faom F5, t in desid to gt only one chisp aput. the (usp o dotined {rom dauugihcaton ploan. The Conmon tchniqus of dajugificaton wsed "oa o Combining the odpt Cerstea a mass ozland Mean ol maimum 9 1akagi -Su qpno Fuggy Model rs Method) The fomad o< the fussy ul ot a Sugno ug34 model in F x is A cnd y is Bwhee A,B aue Vaiabey I and tugy set in the astu cedents ante Z= e4) is a caisp hncton in ine s B THEN ast (Ts model io (onsequent polynonmial dast-oadea Sugeno u3s4 modal 2- 4) oncdea polynomial, 1f f4) is Constant , e get geuo-oAda Suge sbp1 : Fgfyg the input >Fusy indeaence paOI) Wndea Sugeno mocde fusdt ope Aatox tnut be moce! w. the Slm ae made quy - hee , tue p$tep a : Applying the juzy opeaator - he aplied to tif 3-x and 5-y then outpud z= aztbytc Fok aCompaaison blen Mamdani 4 suqeno t odput membeaship dunchon :The main diyeaente bjop them is on the bauig 4otput munbeaship duncton The Suqpno olp mmbeaship dunctio ae eithea inea o (outont COatt Agquga tion and Defuzzificaton Proadua: Method. Ihe dyeence bla them aho lics in the Ccnsequune h334 ues. and due to the Same hctn agguqation and dejuiijcanion procedua also diger 3 Mathemahcal Rule Mou mathe matical ues est Jou the Sugeno ule than the mamdani kule 4Adjustade Paxametes :lhe suqeno cot kollea has mou parameteRs than the Mamdani contolet. Saqero model o 3eAo oxded, the outpt evel z is a Constant b/o HAZ Inpud1 Thputa suqno Model The main .Irpd Munbeiship duncton Znpud membeaship AND Ruu sungth duncion oudped Memben ship iur Z-xtbytc advortages a Mamdani Mehed, 1 ct has wcdespaad acceptance it is wll - Suitable dor human yp Advantags af Suqyno Mcihods. IH (omputahonally eyjicient Hai QIH iu compact and woxk1 wel weitn uniaa tachnique, opimizahon techrique and ada tlive tzchnque 9 Tt in best Sud foA mathematical anay su. 4It has guuanteed continauity o4 the ofp Sujaa, 14.3 Architecture and Operation of FLC System The basic archicecrure of a fuzzy logic controller is shown in Figure 14-2. The principal componencs of an FLC sysem are: afuzzifier, afzy rule base, a fuzy knowledge base, an inference engine and adefuzzifier. It also includes parameters for normalization. When he ourpur from he defuzifier is not acontrol acion for aplant, then the system is afuzzy logic decision system. The fuzzifier present converts the crisp quantities into fuzzy quancities. The fuzy rule base sores the knowBedge about the operation of che process of domain experise. The fuzzy knowledge base stores the knowledge about all che inpur-output fuzzy relationships. It incudes the membership funcions defining the input variables to che fuzzy rule base and the outpur variables to the plant under control. The inference engine is che kernel of an FLC system, and it possess che capability simulate human decisions by performing approximare reasoning to achieve adesied control strategy. The defuzzifier converts che fuzzy quaniies into crisp quantities from an inferred fuzzy control acrion by che inference engine. inpus Normalization input scaling factors Fuzzy knowtedge base HFuziier Fuzzy rule base k(*) Interence k)Defuzzifier (y) engine Normafizalion oulput scaling factors Sensors Flgure 14-2 Basic architecrure of an FLC system. Plant X Outputs Slates The various steps involved in designing a fuzzy logic controller are as follows: IStep 1: Locate the input, outpuc and state variables of the plane under consideration. Step 2: Split che complece universe of discourse spanned by each variable into a number of fuzzy subsets, assigning each with a linguistic label. The subsets include all che clemens in the universe. Step 3: Obtain che membership funcion for each fuzzy subset. Step 4: Asign che fuzzy relationships between che inputs or staes of fuzzy subsets on one side and che oucpucs of fuzzy subsets on ocher side, chereby forming che rule base. Step 5: Choose appropriare scaling factors for che input and ourput variables for normalizing che variables between [0, 1] and [-1, 1] interval. Step 6: Carry out the fuzzifcation process. Step 7: ldentify ehe ourput contributed Erom cach rule using fuzzy approximae reasoning. Step 8: Combine the fuzzy oucputs obained from each rule. | Scep 9: Finally, apply defuzzification to form a crisp output. The above sceps are performed and executed for asimple FLC system. The following design elemens are adopred for designing a general FLC system: 1. Fuzzificaion srategies and che interpretation of afurzifier. 2. Fuzzy knowledge base: normalization of the parameters involved; paritioning of input and ourput spaces; selection of membership functions of a primary fuzy set. 3. Fuzy rue base: selection of input and output variables; source from which fuzzy control rules are to be derived; ypes of fuzzy control rules; compleceness of fuzzy control rules. 4. Decision-making logic: proper definition of fuzzy implicacion; interpretation of conneccive "and"; interpreraion of connective "or"; inference engine. 5. Defuzification strategies and che incerpretation afa defuziet. When all the above five design paramerers are fixed, che FLC system is simple. Based on all his, che features ofa simple FLC system are as follows: fixed and uniform input and output scaling facrors for normalization; fxed and noninteractive rules; fuxed membership funcions; only limited number of rules, which increases exponentialy wich che number of input variables; fixed expertise knowledge; no hierarchical rule strucrure and low-level control. ) Creetc a Radog Toihial stak. (R) Kvalngto titnes. ) Reproduce (Mentathion) 6) let qensah'a. Acfal golutiy Eneoding $pee txcoing pofulaly tautaehorae elGe X lele (lonpikhiau Spate) &Bepi: recle a arog Io lhialae: An do chomoromes ) This is unlike the sikehy fo Skp2: kvazle itnas: Avae tor fitnes ts asiseclt east, solu h(Ciomosowe)depeneing on hou) clox it acually ts to Bolngte poloy (Ans arirtng to tue answe ofAe desired rode). Tese Maslios etG lot to e otused witu "anwers'l to the pblm: tnk of ORdles to Lash the awswen. )ko ; Reprodute (ancl ehilerey Muae): The chuomeoes ith a|nihalpofaton I|Cress ovet ]Mulahoy (emia Yes No Best ndirduas Sutput 43 ngover"): vesaky tolains aeluto tiat chvomotOwe chomoSOe b:otlotIO wrepoveinchvoumaones Finas venetr' ComeRpoveine eewctones DFilne 2) popubhey 6ize Ne 233Fitnes tiners-proportovste seletoy CRoulelte wheel Raulirg Rokte uwnel samplng Fitneu -preportavate elutay Cocer e-bilcowotemes ithke tz lloswne prepethes: (3) 1E B SDse geleded , crevsever s penfenccl. B:0ol|0 E :l01 10100-p:00lLoO0 Bestfit shinp froy prerions peptoy slet ltthe Rtnes ofpopulehn (0 Entaig ( eding) salectoy ) Mutktioy Tables showw the ttnes rale tor the Corekondig ewenpsomes q fisuo shows tae Roulelte ested selerhny (Retonbinay) rts er 5-Gne a eveccle of a) Biay kneodine alowi 4/4 =35 -Knceding is a proe -Thh precss Can be pevted usig bits eay trees/ avays repreeng tnkges ae - he entodinp depevts maioy on solving the pulg. ae. 45 -iofe no. f real nea. can le vepetekd. - No- cf real noi. repreeulecl itvcaer coth ching length. cuveweUSoe, 2This enoing wes shines wede p of octol yor Co-) Comgome 1Jcemaome 2 15423314 () Headecinal kincoding OB46+2\6 -ilhs eueeling yes shig wade up of-hexadeinial nag Co-9, A-) chomosowe 1chvomosome 2)temuttim kntoding (eal 3DBA Evuy cromo some ia shing of nor, Repieendeol n a Reywena Konehines Cone ehons hane to he tlone alter qeresa oeneon is Covplele Ty pemathty enlne eveRy chomeOme aChomeemed pemutM entecing ts dy wetal for ciceng fvreoeme. Krn tttuig plyome types of-cvotsoves Mutr covetev wet e wace to leawe he ciromosone lonnstent(ie;tae cvomeseme 2(e) valve Knoding )Kvey cvowsso we is a|5 | 5 3 2 6 4 4 48 6 5 6 231 4 4anyhie (onneledto the pum: This dne Putes lotstiesult Yaes can Bowe, vew OresoVes Mutetin spette br the pln ehronegome A (. 2324 5.3243 O4556 23213 2:45A comaome BABDJEIFJDHDI ERJFDLD FFEGT 9) Tree Enedng Mahin fol Conveeted to plolmy for m no.}real noR or clakrs Chromne eCek), Chaek), ngat), (komsad) Mett) Fht eneodin is wainly wed ert eyelnp tinmexpreein forgvete (2) Selectto yCutl he eleeted': New populahoy -he ises te ttnex hunehmie beker chanco that ay toeval Me elehm preue is oletrec as die degree tb which he knes oves Beuttewive enesehbns. -There ae 2 dypes of elechioy schewe s -)Proport'on ate -bascd selechts y -eilneavek G)dial -havcol selechoon -froportoake-beccl geleehay peka ou itiv rcuals baecd eyoy ther Atnes aus rekhveo thetthes of the oto itvidd hthepopehoy: taw Fitre butupoy taeir rank withiy Hhe popuetn Weakseleaty will resut y sloo roluchion. (a) Roulette nlheel selecioy -The extetd ane of ay bdiyeuel s todtycal's Ptnes divil cel by beaeel fitsen of te poplady. eh idi utue ]1s aus gned a slice ottthe Rouletehal, tte Sze ofte lice being proportoval tbAhe idiviclel's ftney. Reuletkcwheetis easter toingpemantbut s voEy -the kade of evelutiy deevd onthanananee op hes the popaty Thus techitpe Ranclomy seles a poment Remthe popdethon dirvupive tou Ralek heel alecioy e move Rauk lehoy Banks dae poplahay scNevy Chcwegw -.Fao kers p qeleiay pres e when te folne Nanncc low () Touryawunt selectiovM etournament seletey stae gy proves eletve resuye by hollig a tonaento wpehiho am ly indivtnals a bcst idividnal Rohe tournanets theone wth the ilut ftnes, who is the oiner of Nu: Tourrammn ompeiticvs he winer a e tten ekl into te mating Pool. Thetonament orttang regealcd wl e macng poel for goneseting ew cftpinpsAled. them Pool congpsine the toumaminer hag hige alenge poputay fihes The fra diferente provides the Selechioy pekline wlieh tis Gt to inwe tue Rthes of te Succeclene geres. Ths meted y mone efkrient leads to ay oi ctoy. (e) Boltaann seletoy Cortols tte sede t seleehay atoeling dud wheh qhacally lead'y to ay intraue. ig the slecioy preite ) Stocste oriveral ampling to m preeet-selhecle -prevides Kovo bins c Mininal spraacl. -The indivduek ane weppl to cantiguo Regneh of line suehthateoh indiv idluals sesnt eava in ize to it Atnes eacty a y Roule He hecl selechiey. - Here camaly spaed porsts areplacclover tee ine, dhereare. irdi ytdalg to be seletec' .wanyas ane 80 chilol 2child 1farent 2farent I(6) Teo tont cVoSOves qood tavets,else it severely hages hing Wiag tly te site schose, Leter chdray can lbe doined oy ng i SAnge sLiis nettohe Cros sitg ave eelaed. p Crossovet pointis seleed tardody aovy teley offte etel 4the secfoys affen te cuts cehompeol. tere, acro% pY Taoo maiae, comesome are cut one aesaodie it ces of aking 2parens ions 4) Single-post cooves Rrst doose tanelomy ome Crossoles hore that hcreaes abter olfspring -crDssoes opetor i7 applr'ec ttte mehng pool, oth the ol Producing acilel rom thewm ste roces, of hking CrossOVr CressoVex (O MuHipcirt cossoves ^-poin cress ove ase sdectel Randosy aou acicle c intmeray eehonged. Parent elitl1|Lo child 2C) Utovm eroyoven t can e nohec fat ile poduerne cild l,shenthone Pment 10 nthe mak the qene is cpied frowy fent 2, , han fre pares 2cild 2prodncing lepud fom te parentl. 211 oo lo 11 oe) thvec - RaventvossoVes 1s talkey fotae ofspngtemwe e toit- em, the ttuel eveutH Panent2 farent3 chitd ( shutRle eOsONeN, Befeve Vartabes ae Kandory suftc) to lbott emet After orsbves , he vaideg ite cttspng ane whuled ) PreCedence Presevatve Chossorer CPPx) -Used invehile oting pns 2Selet pawentno-lt/)\2 ) orteved ros oveN citd1 APrentI 4-2 336 5ease seheduling polms tet 31%5 21 irdenctauged. 4cild ChromogOme 4Mulaty pavet cluomosoe Is fliped (o to faud too)eilel generted for 41in mttoy ehromo come, the pading bit obell as fer Raoy clisibuting qenete ihy Mutety plays te vole of re covering he ersenet clill powent parent 756 43i4s 34) Mutatioy 23chilol 2231 45G 2porent I2, 4, 5tPeent (o Rashaly male hed Cros oNer (p<) child 2sent 2C) Revesng A Vanclay tux ihoy ave vevoseol clitd cawomosome s prducd. lchild The, an'ons chowge 6) slall toz lmit-Te 2) Eapsed Tine -The qeett procen will enol oshea sfeatd Aione hey elapol (3) Gm of Fitney 84 t-The GA Sbps when te Speitec) Gonit 4) 84eltoun Fitres Ws-hays bas erelved. Atney Peor Soes ifene it no med te objetvetnehn fera sspuote io tucobletvo kuretay durng an rkel ane