Scribd
Upload
21 views33 pages

Aiml Unit 2 Notes

The document discusses various search algorithms, including Heuristic Search, Recursive Best-First Search (RBFS), and Memory Bounded A* (MBA*), highlighting their features and applications. It also covers concepts related to genetic algorithms, local search, and constraint satisfaction problems (CSP). The focus is on optimizing search strategies and understanding algorithmic behavior in problem-solving contexts.

Uploaded by

Haritha G
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
21 views33 pages

Aiml Unit 2 Notes

The document discusses various search algorithms, including Heuristic Search, Recursive Best-First Search (RBFS), and Memory Bounded A* (MBA*), highlighting their features and applications. It also covers concepts related to genetic algorithms, local search, and constraint satisfaction problems (CSP). The focus is on optimizing search strategies and understanding algorithmic behavior in problem-solving contexts.

Uploaded by

Haritha G
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
You are on page 1/ 33
|e GO UNIT: It Prowiem Soiving. METHODS “HeurBt'e Search -- Tiree Typ, Tematie Duphy a® (100) tam l X RBES + Recursive Best frast Search % Memow Beurded A® (MBA*) (ma*) (ov (sma*) Tlerative Dapenitg A* seanth +. Tpat : ° Ft B Aentlar te HemRe deeping depth - fect Seanch (10-D6 but the APterence ble IpA* ard ID-ppS & thu depth cut-off uso JB f-cost rather than depth only. * The keg feature of Zpa* lgertthm sf does not keep & tock qf, cach vite) node / %E which helps Ih saviag memory consumptin ard can be used shure ‘ts Conbractitter * F- Score B notin’g but the west to reach dhe node ard heurGhe values the node ln) = qin) 1) (path cost) example uF a (i) XB) Geal stote @ @) Furthen - expanstfon taxed on minimum threshload _ $-score value & threshetd & erm Stepa-- To Select the minimum 4- valet ‘Threshold =6 Stepo:- =A FSgsroratur=b, Ahreste 3 grb f-stere value > threshold) festher np expansion Next wminfram threshold value TRresidd =7 =6* f-sore vater =8 Bra futher No expansion atop 3... Next méofmam Ahretreld Value Yvatus (1, 18,12) Thveshold al y) Ciovs) Reach Goal stale =8 ~Swove value =§ 7a>8) pes go can exprheg sucase BB A > EF adc 2 £8. Stop RR proass ) Hl RBFs - Reunsive - Best- frst Search Example & faken from AX and bect- fist ceanch . "a t cet node 1 flih (enpowy | | i as @-© = blew) STEP 1: RBFS (problem , rode, P-Limth) >» RBFS(Protlem. A, 2) Titel flint tale 2 dedard as © | node =A, f-limit = © | | After A*Scax) Successor of A > such a-fS 7,23 ‘ L . 23 My NA best 4 893 (s) minimum Value Now botore Preaade Chuck for tht Went value 393 {@ (Tru) + Now Explore Before that chuck for Hu alternat: node altevnate minimum Value © yt vesult - best-F <— Regs (S, mi (@, 447)) © Rers ($, 47) * eS A2 Altesmativa! minimum felumnit a G @ 343K2(E) — pug(qy HY Now gzplore ¢ | STEP 2: New agouti Tecursive function in RBFS Rips (Syyy7) PE oe node =s f- UPmPt = gy Guusaoy S = Sut Oh = FEARS | Lbb Wis Sub 413 | best HIS CR) mintmum yalut (> | Now before precord ag check fer jhe. { lott valu IBC Ayt (True) by algorithm —> tow explore | node < B limit (explore) | Before that check for the alternate node aliermbe minfmum value + ts alter node +4Is (Second minPrnum ) | yesult best-F RBFS CR, mi (447, ys) | + RBES (R, 415) f-timit a= f-litnit co @ | ¥ be 4 me 49 @ d \ g-lamit byb Gs] juste) 13 (E) Ais {yy 4i3 ; : reach our goal node Memo: Bounded 1 (OFF OPT OG SOLO | pt seath (MAt Ox) Sna*) Cord Hong : ) > 28 similar te A* but restrict the, Memory ste to fix. Sto Ane avatlatle memos . > Whin memoy a8 full St -connot+ add 0. neato nile to the Gah 1 > Without drorpiig on od one .Crode) 2) To add a new ride alwcays drops The Word - laf rede (Ste ene sath tia hifest F-velue ) 2) SMa* expands Aeast heal rede and delites the cenrst- Jah node Memory $ Pao *8 (0) =g(n)+ hin) Yoot node Oo node Op 4.- as) 0 “ . gin)= 8 0 : we & : — Stay 0 3 _ bin) ain)-0 4 @: ah? $7Bs3 ~ Ly hig) =o Wort teal node Spas nede She Joxgelho dle Value sh the porathes ¢ Sorgotien node wstttin 8 Porathesd, . Srupdale sith Value uotth nts F-Vatue feason:- ———updatiig § -- | Compare the fev aten node cand B's) B79" need te cad fhe. sh the parathests Bb consider % % most promisthg node. ( ie me mode 8% this 5 Now oe ned ts Jtpate he pom [zea] © update [S=4] mr?roum value Bb c=6 Zo 10 stp 4:- ng ( He) , SIG ar ytree hin) =o eachiba tre goal ahh maxima cost So we real to soconstruct the yuo Reah “the geal pode with { Value <6, $0, noe nacd to pat the {-votu provttus node aofth G (le, Azb) ‘But Lis Ve of ba 5 (ie, B=s) compare to the 6% s bast one go the graph f above os ¢ Ra Py Peet role BRS bak ABE & remove A from th memory. MCSD 3 Kis) &(5) Ob WEN (etFnfrati) x b: 6®e A Reconstruct : - Ne nud to update C26 because 9 fogten value & also 6. { ial path oS (32826 Ff Here seach the gel nede with te Value eo} 5. compare te the dorgetter | node (6), 542 bottor . | htneo toca searet ALGORITHM AND Optimization PROBLEMS : | | Loca SeAeH ALgORTHM | at ot | | Hi | Simulated | |Jocal beam] | genehe | eg) j | & h Algorithm | || Sc? | | Ancualetg | \ ene ¢ . v +L. steeped asunt . 7 ke Be : | La gkchaste HU ARnbag | ah i | LS fast chefee bill clfinb z TA arden Restark AFM eltmbiag al oats. wi Genetic Algorithm > Populakty 2 {fires > sclectin > Orssover => mutation POpoLATION. ‘ Jn tesm e Biology +> celleckin ¢ tA | In term g Algorithm collection of aks FTiness Ps ay jargq Biology > More healthy , less Prone te disease Algorithms closest ts the foal Sdubtn . | Biology - - = _ Slettfn:. Lt ee 4 Algo > bfolegr ly ft - Ly sclectutg staler Hat ae doest % the | | Seluhian CePinadk ) frat oak Lx most (oss Over - Reprelacig Biolegy —> Maki or Valats Sti Stet stats | Algo thm > cnterchangen§ | varittion Bteagy > change Bilore > AD Algorithm > Alzeralito Sins As Gr components ) BQuun problem | ty prrorge 2 Quunt 8 sid ess beard FP Such | | o ung that ne guun altacks cach cther . | & i| ica “Mii i i] fara fot = PH EG sty Z Quun preblem wig He genet Aptis. Sep. Represeriaig irdi?duals Step G envatirg an ‘tritfel popslaivin Sheps: Applyig 2 {fires force | Steps: Selechhg Parents for raha ‘to thefr fret” Te - af -a.ceerdanu SEDs. crossover parents fo produc nue geneotin 6:. Mutatin & nue gencoti T 4g ste 1 Repeat until solution Be yeat +: Br®e dhagsh | te bug Aivenstiy 7 bad Podfuiduals ep ae Represeritang ee Genaate (nitOy eet Genaale vanlan avrangemink of § Quacns cn a td chess ( B x Le sisise) (SPISIN BPI) .PHMWBIEIy] ola Slepa:- Apply fitness funclién : Fin = No- d- non attacker pats Quun 1b | Qum a: 5 Aunn sy | Quen ye S Quuns'. | Quune:- Quen T° Quen & : Mt Qe sles | | Trdivduals | FP ness i Friness Yo poy.23 = 24/0 18 a3 | : —Aa%/o | rd rutdual a= T x imme wo gare htt i 9st2y $l $202 18 B ¢ 2 0 ab. 3° ou Im lyse 20 2 26% yeah, Beh Mo ae B1°%/o (yee ae > Blip Yio dolectian / eer various mnethids “ Soloed Y There ee 4 Poutetfe whl, Tournoroent Untvewal Gainplibg (605) ¥ populalfer yi Atytdod ona onl actondihg ln tno avsped fue perunl ag? filoes and bun fond Pook ase placed, prank , Ve i fe ahonsie whol JS Spun and those PrP vitladds oot ° folecled fl aohith the acd potth, ce Parte whi tho Wheel Stops 1 yor Bi Boy — opus Di h.042 4— A 24+0%/o Band Dar Soleled a parents ala} [sie fy ry ; [sob [spi [ofa | H Repeat » - All Steps axe reprated unk best Solution uF reached % Bast Solution = Hijfest ffiness Score (as R 4nd @se) - Quuns | No- & Mon atrackiig quan | a | a. Q. | 6 3 5 Oy 4 | Qs 3 Qe 2 | Qs \ | | Qs 7 | Total | 2h Lo aa BW dirbfg alyatm - 2 0) 7 belve te leeks wld problem - c-1 CA] pitt 9 Oo. ee Tae ¢ 41 fe} B+) {35 | oF i £1 ) {| Sloat Gen), | 1 fr ech Mak ste 3 ret on te ta | Sappaatto be este the. | - "fox cach block tha ub restdg ma the. sor00g thi'g | based on geal state Value uf given - | Natu fox the Slot Stale. B7-! | a r Ca41 ° -)+)41-1= | 5 fe 4 D+ ‘ : ° jee | anes 7 Vatus fer The Goal state Dat CH, BSH, AP! Flaptn fp -' ¢ t) [ + (> of uP OL TE - fel -ferA+ elas? Stoxt ro DI-! B as ahs (0) lel 9 " to) o \ arate Eas & wwe move foo giebal Fecause + getty seco Himes ° fn fiop mover in local Aewrdh% furtt : m (OR beth Ty awe so He endad wit) Joal maxSum cry teal opin So tah at a ddyantage of een} marfinam | bo aed aeve on fe Tie gre hems ee gieend = Peri Janel eek Bick tint hes soneng Support bolt every bieek aR fhe depprt Shulire ‘ ae . f = eneeS=' GY ¢& | Sater B BOR m 1 j 2 S£=2 j Edian ¢ 32 reds Sf | = Js 2P=S S — & 5 MIBRP=S i Belen D © S ) ) erica - ‘ey of #| [alo [p]o Gool stati (ante Constraint, Salisffeatton problem ( ¢sp) ¥ Cop ave mathematical Cus libn Arent Jona se of all obfec6, sshose state rust soliify | tl no. constraints OF lemttalsns. 4 A represents obfecls fnate constvatnl, Over constr aunt dalibffealen methels wa Problem as 4 | wlledtfon Varfablee snitch 4 | Golved by é + An assignment D complete even VarPall re Austra Oo Value | # A Solukdn fy Cp ub a Complete ausfgnres | | | | | | that Satisffes all conshratnts i | \ | The Appltatibn : i * map Cblour,2 “8 | ee Problern ‘ dna. rawey Frterprctats, ae Problem flocs Plannig for MIST yo Formal defenfifoo of csp: - A CSP consc8t of + 4 footie Sh Vertalls yi, fn | ¥ pon empty domak pess?ble Values Di D2, 6. Da awhere DOe AVE VK Y fotte Seb de conshsaunts C1, C2,°> [mets eRe Values Hhak Vari fe cath Varfable. etm ~ gach tonsa cf ables can fats, 0-4-2 Xf M2 to 4A stale & defined ox 80 assigoment of Value | gyme oF al Vartables - (A consistent ass ment does not Violate the constr’ B called as Consistent or legal assynmenk j Exargle a, CSP - Map celeéq problem | Tre protem te color the Yegfere as given map gach that no joo cdfacmt Yepitns hove jhe Same unables are Te Celovs - | & The egitns af te map ave the get oh pesstble Celers fer 16g iB tix devad we > Weslo, i te 2 cag RANA, Representation at Problem * - Me fase 4Ottory, SA > ðAastia — oo K lovable. Yeo TWAINT ONSW. Ve sa, Ty & Domats Dr = $red green 1 Bee § | Constraints cadfaent egFons must have Ae erent tolors eg: WR ANT Xoo, want) mw be ub {hred, green), g Cred , btet) efgrced ed), § °, o. ck Salut Oe Cerpleta ard canststent assignments Qe Wa » red , ( ms Nt. queen) yo Nn ¢ a= ved fwa \o Ngive ren ved a ye ved vt Ge ble, Ts green Re be OOO eripe Anime problem sel csp Rules {or aripk antihmeti prctiem :- o- onl 7 Rr @ DRL wings fsorm he 02 @ Fah Vastable sheatd have OY Suk Value . ‘ ARE Value a agit oats R3 @ pach Syméel Hepes Ong put The preblem . ‘ Py @w. {A the unlus of leltes uh esp gs @ Reh may be only ene Setutidn te the * Preble R-b @ he numer base unless spate 8 10 o.7 @ Te numbers must not begih hn Fer Frample t- SEND + MOEE = Money wafabies BE L DENN, OR SVS Paness fo 84 | Confvatht —Pule (addon) = &y: DtE = if Now assign Carey fo each column . Ne Ls - Puample Ce G00 : step 1 Re the column > | carry s a sutthont PY xesulk a M hos fe Value f - yortable tutes but ik has go Rk stleare hak Cs (NO valuss ts add) Sep 2: | const vaurk Axed from. Cy. | we have alveacky acumed (oe has Value *t! | [x4 means column 4 yosult have Should be qreoter ox equal to '0 i} Oy 4 Ste & jigfe (y + S+M= 0410] le know that Mel Now Cq Vatue Can ether & ler) (1) | =o OFS MO | | x0 | | Cyet | | l4StM= O Choose8 eS etna Cy =O cq S4tM = OF10 oe ae EV jot as us assume Seq Med diy = toro lo to Oo (ose ] os Tae), ES) Ge] Skip 3 gonhradt dered prem Cs | (er Eto2N t | goad C9 Vale, can cfther 0 (or) | | (gro > GrEA ON ovedo aN e (o») Gee cu, Ose | Cart [PEO oN TW al nol poss? ble | So wwe Choosth laEtoeN Wo know Jhat yatue & <0 WEtoeN | (weeny Wang 3 conser deveitt pom ca e x Cot Neha & eRe, ca ak | Lb coe NER & FTW —® Raw Sub Qqactid OG) Nertt ad the above cqualite vy af @) (2a Cor Nak FEO be o for) I eihe assume the Ca Value can ef they chase ey OENAR SEED 2 > Nak FAto PH1eRs Belo not possible beceus ¢ oa not“ unfgnce Value) ° So the asseumption f OVO | So cheese. | ¢: Cot NE Rs EXIO SP IENER = EHO Fltetl aes 2 el0 PF QIR=ID (& we have Valu Constryacit derived Prom One Ter af Prifal et a Sct to CrADtE =Y OrDEEs¥ 7ewy Diezy feo] Bee Ca Vatte Uf 1 there exite & OMY Rale Nam by : ; Dae» Yato PPerndirkng nds B45 br: | (Poy ta, 11 Me ts, Ib Sep b POPE Fre 8 conptirmed thal co Value 1, % dhe Value for cr dhout) be 19, 15) te, 16,1 t00) 17] | & Charge Ae Vatua ef cd, © Such Nat the result toll be fanny Assuméig Des Ee ] DIE = YAlo | Woe gato { : Value, 5e1T = 2410 | by canny | & Bey lend | | from step & eEPI" y GQ) Neetl : Sub €=7 Neo 4H ‘ ‘ Nee Nea Wak not possibla above assampifon Step we aon because P=& S that Now asscuntig Ded and €F> vat (48a The Vals. fo be 12) Dr Es yao Atty = carry te) = Y410 ey to Netad Sub bes Ne tt [hee] So Nhe. Vals ve New| [pal RB Eq) Neer! Nest \. N= Cte does not forplite swith other alphabels) AGel Solutfon Jace the alphabets port other dectmal . | Rep pritfolly 2 1 ou) fo (1) Co) Prifally sere 1 | ts | Sq) E(s) Niy Dia) | may Oc) "ry Fes) pe ht Por 9 te) My %@ No Fo Ze | MIN - Max PRoceDUee Tie TAC Toe problem | AT uf the pony to yird yw. machine “6 yhudk tke a hernan ard sofa the game . \ Hew we ane. gor fo Make, = tach plage s marks & a, op resenlakie, 3x2 qld by | sy’ and Yo? wy dum. | Tk wan tuo payer come fq:- Tic Tac Toe PROBLEM . | ol o|% . \| x oO > urrent pasttion 5 \ mark player / ~ . | (x's move) ote x ath su : Joss = (0) o lolx K o x oO\% ‘0 roan | {o) | Apna - BETA CHONG) Cadyenceet PRONNING preduilien OP © maa) max the number aplored ential thet ‘east pa ge To Reduces the 70° 4 nodes explora » prunnig a done (TAS B done by aug branch Source Strategy worth cBetudes eve Bounds 4 a Is fo each player thes shrategs ae called “% Alpha. - beta prunntig bony sty vedleeee OF Tn mfn- ann FE aPewoases expen” wb the clepth - f | YBa kechnPque- wht — helps nedes genurator > wee dower bovine) on TH Value hel maxtmtgeqg the node may be assiyned _B = upper pound on fhe Value that minfinPy Be. he node, may be | asf necl (vedueey the tine requires] er Seach ) \ Foow 4 y || Slept: - | Taitially Sek Y= - © | pet | |B follwss Das geonh tere ch®d, no Value LR the | xe =-0 and Bete | | teptchilol move to left node 5 Peete pat ena tae) thy [eft most node it 4-2 move To | mfr Vatu. te 3) da%e The mihimumn Value hede u® level 1. echnfqul + 5° (mar = change « nih + charge B move te the BE again Sufee left most Valuo, Be -3 favo aofll assign -$ Search for | node Fd fle is7-3 ‘Gu value e--© and BP Re. 3<(4+@)) fe + C3<4+@) © mi ~b ng (16 4 -5 a (2 into next ferminal Aga Saanch fc mimam value PE ard check So leave newt hecle. Ca] Qr-3 & aga’ assign the Yalu ts the provides node -3 77 | wu ols Be-# 3 4 go 38 aufice Hue other nedes ch fie left most rode leva t| ave empty 1 the Value go the valux of af pared “to the root rede o& changes - (88e Sep g.- CpiReT node vistted) -3 B greater than -©) i ( vs “b (wo, qos FADS Tru, -ypamed . Here ¢ 223 and Bs- -2 Green pe-s (10 cack the rezk node cas chuck gard % : Here vB ltorditim tau) no nad f chet 32-5 Other neces Step y-- Fish vd Send rede & vit fos we move to tied rede - mat - | {ei Solulion pm Aim ne } Ca “a 3 le 7 7 cul Off a) max (f To solve the Jive {| Aon. anh mh tons mee A alaorithrn ard also | B Nipra- Beta pruncig . mark : s) @® @® @ =i lara |) aa Ye posstbplrives:- maxe[a.g,-0} | mblotted | man's [o,4,-03 | mh La, ote | mere tier =) man [5,0 -°J mox=fo,-1, -@] By -tathy Npha« Beta — prantoy . mat ee BA Yo Pap) 5 (L Besa Beye po Cif py Gyo7x0 8070 rudd lo chary. wv true Step hy 00 cvlpk Arithmett’c puoblera wnt op CY cross + Rone = DANGER co ty o ¢ 5 Cay Ro s° ic) oO ue tw Rye A oa D AwyN g, FRw DB wy Del | Jet us assume f=? , So Rey fot ws agaume C7 $0 he grad E> oa ReromeRy vatus ave 8, 16,5 4 ape cusume O-sit wll bo false . ° Suppose 0 WwW equal to § and C3 o> Bty = 12. not posibe 7 equal to 1 and) Cy -y ft = ct not possible e a equal to 6 ard ty b1y fo nat Pox? ble O Wequal to 5 — and Cyr G44 = 4 not Possible So tt cause, w00 base hange th ‘ y YO Marg he AXcuMp lay alvead y ass ned 9. [Je Vie ULM ey y OY CO¥ [yO fr (8) 65 Pegg heey’ yo 1) As Ke Me Vas Sof ty Rarwiihidg tea, Og” 8;'4) 44 Usigy 07% te C7 FA b+ = 16 Hod possible bey D yy 61 21% 1 Posie 6:4 . % (yy Abs ly not Possible | Se cessed 0-7 Se Cpe 5 28 possible | fa LBhFOo

You might also like