CSE 573: Artificial Intelligence Autumn Search thru a. Goal Based Agents 9/28/2012. Agent vs. Environment. Example: N Queens

Similar documents
CS 188: Artificial Intelligence Fall Announcements

CS 188: Artificial Intelligence Spring Announcements

Announcements. CS 188: Artificial Intelligence Fall Today. Reflex Agents. Goal Based Agents. Search Problems

CS 188: Artificial Intelligence Fall Announcements

Outline. CSE 473: Artificial Intelligence Spring Types of Agents

Announcements. CS 188: Artificial Intelligence Fall Reflex Agents. Today. Goal Based Agents. Search Problems

Announcements. CS 188: Artificial Intelligence Spring More Announcements. Today. From Last Time: Reflex Agents.

CS 188: Artificial Intelligence Fall 2011

CS 188: Artificial Intelligence Spring 2009

CS 188: Artificial Intelligence Spring Announcements

CS 188: Artificial Intelligence Spring Announcements

CS 188: Artificial Intelligence

Reminder. CS 188: Artificial Intelligence. A reflex agent for pacman. Reflex Agent. A reflex agent for pacman (3) A reflex agent for pacman (2)

Today. CS 232: Ar)ficial Intelligence. Search. Agents that Plan. September 3 rd, 2015 Search Problems. Uninformed Search Methods

Course Logistics Textbook: Artificial Intelligence: A Modern Approach, Russell and Norvig (3 rd ed) Topics

A search problem. Formalizing a search problem. Our Search Problem. Our Search Problem. Overview

Searching: Deterministic single-agent

Search: Cost & Heuristics

Announcements. CS 188: Artificial Intelligence Fall Office hours, Section. Today. DFS and BFS. Recap: Search. Lecture 3: A* Search 9/3/2009

Announcements. CS 188: Artificial Intelligence. Costs on Actions. Recap: Search. Lecture 3: A* Search

Self-Adjusting Top Trees

Exam 2 Solutions. Jonathan Turner 4/2/2012. CS 542 Advanced Data Structures and Algorithms

176 5 t h Fl oo r. 337 P o ly me r Ma te ri al s

T h e C S E T I P r o j e c t

Weighted Graphs. Weighted graphs may be either directed or undirected.

Easy Steps to build a part number... Tri-Start Series III CF P

10.3 The Quadratic Formula

Trade Patterns, Production networks, and Trade and employment in the Asia-US region

P a g e 3 6 of R e p o r t P B 4 / 0 9

Problem solving by search

Tangram Fractions Overview: Students will analyze standard and nonstandard

P a g e 5 1 of R e p o r t P B 4 / 0 9

(2) If we multiplied a row of B by λ, then the value is also multiplied by λ(here lambda could be 0). namely

Planar Upward Drawings

RUTH. land_of_israel: the *country *which God gave to his people in the *Old_Testament. [*map # 2]

4.1 Interval Scheduling. Chapter 4. Greedy Algorithms. Interval Scheduling: Greedy Algorithms. Interval Scheduling. Interval scheduling.


1 Introduction to Modulo 7 Arithmetic

NUCON NRNON CONRNC ON CURRN RN N CHNOOGY, 011 oo uul o w ul x ol volv y y oll. y ov,., - o lo ll vy ul o Mo l u v ul (G) v Gl vlu oll. u 3- [11]. 000

-Z ONGRE::IONAL ACTION ON FY 1987 SUPPLEMENTAL 1/1

d e c b a d c b a d e c b a a c a d c c e b

Appendix. In the absence of default risk, the benefit of the tax shield due to debt financing by the firm is 1 C E C

Who is this Great Team? Nickname. Strangest Gift/Friend. Hometown. Best Teacher. Hobby. Travel Destination. 8 G People, Places & Possibilities

Series III, TV Breakaway Fail-Safe Connectors Quick-Disconnect with an Axial Pull of Lanyard

KEB INVERTER L1 L2 L3 FLC - RELAY 1 COMMON I1 - APPROACH CLOSE 0V - DIGITAL COMMON FLA - RELAY 1 N.O. AN1+ - ANALOG 1 (+) CRF - +10V OUTPUT

Multicast routing algorithm based on Extended Simulated Annealing Algorithm

CSE 373: AVL trees. Warmup: Warmup. Interlude: Exploring the balance invariant. AVL Trees: Invariants. AVL tree invariants review

F102 1/4 AMP +240 VDC SEE FIGURE 5-14 FILAMENT AND OVEN CKTS BLU J811 BREAK-IN TB103 TO S103 TRANSMITTER ASSOCIATED CAL OFF FUNCTION NOTE 2 STANDBY

OpenMx Matrices and Operators

I N A C O M P L E X W O R L D

COMP108 Algorithmic Foundations

A L A BA M A L A W R E V IE W

I M P O R T A N T S A F E T Y I N S T R U C T I O N S W h e n u s i n g t h i s e l e c t r o n i c d e v i c e, b a s i c p r e c a u t i o n s s h o

learning objectives learn what graphs are in mathematical terms learn how to represent graphs in computers learn about typical graph algorithms

Ash Wednesday. First Introit thing. * Dómi- nos. di- di- nos, tú- ré- spi- Ps. ne. Dó- mi- Sál- vum. intra-vé-runt. Gló- ri-

/99 $10.00 (c) 1999 IEEE

PLAYGROUND SALE Take up to 40% off. Plus FREE equipment * with select purchase DETAILS INSIDE

Paths. Connectivity. Euler and Hamilton Paths. Planar graphs.

( ) ( ) ( ) 0. Conservation of Energy & Poynting Theorem. From Maxwell s equations we have. M t. From above it can be shown (HW)

Chapter 6 Perturbation theory

Winnie flies again. Winnie s Song. hat. A big tall hat Ten long toes A black magic wand A long red nose. nose. She s Winnie Winnie the Witch.

Improving Union. Implementation. Union-by-size Code. Union-by-Size Find Analysis. Path Compression! Improving Find find(e)

CSE 373: More on graphs; DFS and BFS. Michael Lee Wednesday, Feb 14, 2018

Spanning Trees. BFS, DFS spanning tree Minimum spanning tree. March 28, 2018 Cinda Heeren / Geoffrey Tien 1

Three Phase Asymmetrical Load Flow for Four-Wire Distribution Networks

ECE COMBINATIONAL BUILDING BLOCKS - INVEST 13 DECODERS AND ENCODERS

Lecture 20: Minimum Spanning Trees (CLRS 23)

An undirected graph G = (V, E) V a set of vertices E a set of unordered edges (v,w) where v, w in V

THIS PAGE DECLASSIFIED IAW E

Helping Kids Prepare For Life (253)

L...,,...lllM" l)-""" Si_...,...

b. How many ternary words of length 23 with eight 0 s, nine 1 s and six 2 s?

Exhibit 2-9/30/15 Invoice Filing Page 1841 of Page 3660 Docket No

Designing A Uniformly Loaded Arch Or Cable

Problem 1. Solution: = show that for a constant number of particles: c and V. a) Using the definitions of P

Daily Skill Practice

Module graph.py. 1 Introduction. 2 Graph basics. 3 Module graph.py. 3.1 Objects. CS 231 Naomi Nishimura

Housing Market Monitor

OH BOY! Story. N a r r a t iv e a n d o bj e c t s th ea t e r Fo r a l l a g e s, fr o m th e a ge of 9

Stable Matching for Spectrum Market with Guaranteed Minimum Requirement

Table of C on t en t s Global Campus 21 in N umbe r s R e g ional Capac it y D e v e lopme nt in E-L e ar ning Structure a n d C o m p o n en ts R ea

Gavilan JCCD Trustee Areas Plan Adopted November 10, 2015

Classical Theory of Fourier Series : Demystified and Generalised VIVEK V. RANE. The Institute of Science, 15, Madam Cama Road, Mumbai

ENGO 431 Analytical Photogrammetry

February 12 th December 2018

Functions and Graphs 1. (a) (b) (c) (f) (e) (d) 2. (a) (b) (c) (d)

Executive Committee and Officers ( )

sin sin 1 d r d Ae r 2

Nefertiti. Echoes of. Regal components evoke visions of the past MULTIPLE STITCHES. designed by Helena Tang-Lim

CS 241 Analysis of Algorithms

Outline. Binary Tree

Call for Applications

and the ANAVETS Unit Portage Ave, Winnipeg, Manitoba, Canada May 23 to May E L IBSF

CATAVASII LA NAȘTEREA DOMNULUI DUMNEZEU ȘI MÂNTUITORULUI NOSTRU, IISUS HRISTOS. CÂNTAREA I-A. Ήχος Πα. to os se e e na aș te e e slă ă ă vi i i i i

INFLUENCE OF ANTICLIMBING DEVICE ON THE VARIATION OF LOADS ON WHEELS IN DIESEL ELECTRIC 4000 HP

Applications of these ideas. CS514: Intermediate Course in Operating Systems. Problem: Pictorial version. 2PC is a good match! Issues?

The University of Sydney MATH 2009

Theory of Spatial Problems

(Minimum) Spanning Trees

BACKFILLED 6" MIN TRENCH BOTT OF CONT FTG PROVIDE SLEEVE 1" CLR AROUND PIPE - TYP BOTTOM OF TRENCH PIPE SHALL NOT EXTEND BELOW THIS LINE

Transcription:

CE 573: Atiiil Intllign Autumn 0 Intoution & Dn Wl Wit slis om Dn Klin, tut Russll, Anw Moo, Luk Zttlmoy Agnt vs. Envionmnt An gnt is n ntity tt ivs n ts. A tionl gnt slts tions tt mximiz its utility untion. Ctistis o t ts, nvionmnt, n tion s itt tnius o slting tionl tions. Agnt nsos? Atutos Pts Ations Env vionmnt Pln Ask wt i ol Bs Agnts Disions s on (yotsiz) onsuns o tions Must v mol o ow t wol volvs in sons to tions At on ow t wol WOULD BE tu Polm / tt Inut: t o stts Otos [n osts] tt stt ol stt [tst] Outut: Pt: stt stt stisying gol tst [My ui sotst t] [omtims just n stt ssing tst] Exml: N uns Min Lning : it ul iiny Inut: t o stts Otos [n osts] tt stt ol stt (tst) Dist Dt Pit MP mg ylins islmnt osow wigt ltion moly mk goo 4 low low low ig 75to78 si 6 mium mium mium mium 70to74 mi 4 mium mium mium low 75to78 uo 8 ig ig ig low 70to74 mi 6 mium mium mium mium 70to74 mi 4 low mium low mium 70to74 si 4 low mium low low 70to74 si 8 ig ig ig low 75to78 mi : : : : : : : : : : : : : : : : : : : : : : : : 8 ig ig ig low 70to74 mi goo 8ig mium ig ig 79to83 mi 8 ig ig ig low 75to78 mi goo 4 low low low low 79to83 mi 6 mium mium mium ig 75to78 mi goo 4 mium low low low 79to83 mi goo 4 low low mium ig 79to83 mi 8 ig ig ig low 70to74 mi goo 4 low mium low mium 75to78 uo 5 mium mium mium mium 75to78 uo Outut Y N to in Hyotsis : X : X Y

Hyotss: ision ts : X Y tu o Dision Ts E intnl no tsts n ttiut x i E n ssigns n ttiut vlu x i =v E l ssigns lss y To lssiy inut x? tvs t t om oot to l, outut t ll y Cylins 3 4 5 6 8 goo Mk Hosow mi si uo low m ig goo goo goo 8 Mtos Blin Dt ist s Bt ist s Ittiv ning s Uniom ost s Lol Inom Constint tistion Avsy tt s tt s g: E no is stt T susso untion is snt y s Egs my ll wit osts W n ly uil tis g in mmoy (so w on t) Riiulously tiny s g o tiny s olm tt izs? Ts Polm: Et ll o t oo Pmn ositions: 0 x = 0 Pmn ing: u, own, lt, igt Foo Count: 30 ost ositions: N,.0 E,.0 A s t: tt stt t t oot no Ciln oson to sussos Nos ontin stts, oson to PLAN to tos stts Egs ll wit tions n osts Fo most olms, w n nv tully uil t wol t

Exml: T tt s vs. Ts tt : Wt is t s t? W onstut ot on mn n w onstut s littl s ossil. E NODE in in t s t nots n nti PATH in t olm g. tts vs. Nos Nos in stt s gs olm stts Rsnt n stt stt o t wol Hv sussos, n gol / non-gol, v multil ssos Nos in s ts lns Rsnt ln (sun o tions) wi sults in t no s stt Hv olm stt n on nt, t lngt, t & ost T sm olm stt my iv y multil s t nos T Nos Polm tts Pnt Dt 5 Ation No Dt 6 Builing Ts : Exn out ossil lns Mintin ing o unxn lns Ty to xn s w t nos s ossil nl T Rviw: Dt Fist Imotnt is: Fing Exnsion Exlotion sttgy Dtil suoo is in t ook! ttgy: xn st no ist Imlmnttion: Fing is LIFO uu ( stk) Min ustion: wi ing nos to xlo? 3

Rviw: Dt Fist Rviw: Bt Fist Exnsion oing: (,,,,,,,,,,,,,,) ttgy: xn sllowst no ist Imlmnttion: Fing is FIFO uu Rviw: Bt Fist Algoitm Potis Exnsion o: (,,,,,,,,,,,,,,,,,,,,,,) Comlt? unt to in solution i on xists? Otiml? unt to in t lst ost t? Tim omlxity? omlxity? Vils: Tis n Num o stts in t olm T mximum ning to B (t mximum num o sussos o stt) C* Cost o lst ost solution Dt o t sllowst solution m Mx t o t s t DF DF Algoitm Comlt Otiml Tim DF Dt Fist N N No No O(B Ininit LMAX ) O(LMAX) Ininit no nos nos TART m tis Ininit ts mk DF inomlt How n w ix tis? Ck nw nos ginst t om Ininit s ss still olm OAL m nos Algoitm Comlt Otiml Tim DF w/ Pt Cking Y i init N O( m ) O(m) * O g s nxt ltu. 4

BF Algoitm Comlt Otiml Tim DF w/ Pt Cking Y N O( m ) O(m) BF Y Y O( ) O( ) Ext Wok? Filu to tt t stts n us xonntilly mo wok (wy?) tis no nos nos nos m nos In BF, o xml, w souln t ot xning t il nos (wy?) Vy siml ix: nv xn stt ty twi Cn tis wk omltnss? Wy o wy not? How out otimlity? Wy o wy not? om Hints s is lmost lwys tt tn t s (wn not?) Imlmnt you los list s it o st! Nos ontully ts, ut tt to snt wit stt, ost, lst tion, n n to t nt no Mmoy Limittion? uos: 4 Hz CPU 6 B min mmoy 00 instutions / xnsion 5 yts / no 400,000 xnsions / s Mmoy ill in 300 s 5 min 5

Comisons Wn will BF outom DF? Wn will DF outom BF? Ittiv Dning Ittiv ning uss DF s suoutin:. Do DF wi only ss o ts o lngt o lss.. I il, o DF wi only ss ts o lngt o lss. 3. I il, o DF wi only ss ts o lngt 3 o lss..n so on. Algoitm Comlt Otiml Tim DF w/ Pt Cking Y N O( m ) O(m) BF Y Y O( ) O( ) ID Y Y O( ) O() Cost o Ittiv Dning tio ID to DF 3 3 5.5 0. 5.08 00.0 33 8 Puzzl xx Ruik s 5 Puzzl 3x3x3 Ruik s 4 Puzzl Assuming 0M nos/s & suiint mmoy BF Nos Tim 0 5.0 s 0 6. s 0 3 6 ys 0 9 68k ys 0 5 B ys Wy t in? Ruik s ig ning to 5 uzzl s gt t Mx 8x It. D. Nos Tim 0 5.0 s 0 6. s 0 7 0k ys 0 0 574k ys 0 37 0 3 ys # o ulits li t om Ri Ko snttion Wn to Us Ittiv Dning Costs on Ations N uns? 3 8 9 8 3 OAL TART 5 4 4 Noti tt BF ins t sotst t in tms o num o tnsitions. It os not in t lst-ost t. Dnil. Wl 35 6

Uniom Cost Uniom Cost Exn st no ist: Fing is ioity uu TART 3 8 3 9 8 4 4 5 OAL Exnsion o: (,,,,,,,,,) Cost ontous 8 3 9 8 5 0 3 9 4 5 7 6 6 3 7 8 0.us(ky, vlu).o() o() Pioity uu Rs A ioity uu is t stutu in wi you n inst n tiv (ky, vlu) is wit t ollowing otions: insts (ky, vlu) into t uu. tuns t ky wit t lowst vlu, n movs it om t uu. You n s ky s ioity y using it gin Unlik gul uu, instions n t onstnt tim, usully O(log n) W ll n ioity uus o ost-snsitiv s mtos Uniom Cost Algoitm Comlt Otiml Tim DF w/ Pt Cking Y N O( m ) O(m) BF Y Y O( ) O( ) UC Y* Y O( C*/ ) O( C*/ ) C*/ tis Uniom Cost Issus Uniom Cost: P-Mn Rmm: xlos insing ost ontous T goo: UC is omlt n otiml! 3 Cost o o tion Exlos ll o t stts, ut on T : Exlos otions in vy ition No inomtion out gol lotion tt ol 7

Exonntils Evyw Huistis I tink w going to n stong onky 43 44 Huistis Huistis Any stimt o ow los stt is to gol Dsign o tiul s olm 0 5. Exmls: Mnttn istn, Eulin istn Bst Fist / y Exn losst no ist: Fing is ioity uu Bst Fist / y Exn t no tt sms losst Wt n go wong? 8

y Bst Fist / y Exn t no tt sms losst A ommon s: Bst-ist tks you stigt to t (wong) gol stt B A gol Wost-s: lik ly- gui DF in t wost s Cn xlo vyting Cn gt stuk in loos i no yl king Wt n go wong? Lik DF in omltnss (init stts w/ yl king) Bst Fist y Algoitm Comlt Otiml Tim y Bst-Fist Y* N O( m ) O( m ) m Wt o w n to o to mk it omlt? Cn w mk it otiml? A* Ht, Nilsson & Rl 968 Bst ist s wit (n) = g(n) + (n) g(n) = sum o osts om stt to n (n) = stimt o lowst ost t n gol (gol) = 0 I (n) is missil n monotoni tn A* is otiml } Dtil stt Euon Exml Wn o w k o gols? Wn ing to uu? Wn moving om uu? n 54 9

A* Exml A* Exml 55 56 A* Exml A* Exml 57 58 A* Exml A* Exml 59 60 0

Otimlity o A* Otimlity Continu 6 6 Pos Cons A* ummy Ittiv-Dning A* Lik ittiv-ning t-ist, ut... Dt oun moii to n -limit tt wit -limit = (stt) Pun ny no i (no) > -limit Nxt -limit = min-ost o ny no un FL=5 FL= 63 64 IDA* Anlysis Comlt & Otiml (l A*) usg t o solution E ittion is DF - no ioity uu! # nos xn ltiv to A* Dns on # uniu vlus o uisti untion In 8 uzzl: w vlus los to # A* xns In tvling slsmn: vlu otn uniu +++n = O(n ) w n=nos A* xns i n is too ig o min mmoy, n is too long to wit! nts ulit nos in yli gs Fogtulnss A* us xonntil mmoy How mu os IDA* us? Duing un? In twn uns? 65 Dnil. Wl 66

MA* Us ll vill mmoy tt lik A* Wn mmoy is ull Es no wit igst -vlu Fist, ku nt wit tis -vlu o nt knows ost-oun on st il Altntiv Ao to Finit Mmoy Otimlity is ni to v, ut Dnil. Wl 67 Dnil. Wl 68 Dt-Fist Bn & Boun ingl DF s uss lin s K tk o st solution so I (n) = g(n)+(n) ost(st-soln) Tn un n Ruis Finit s t, o oo u oun on solution ost nts ulit nos in yli gs At om Ri Ko snttion 69 Bm I Bst ist ut only k N st itms on ioity uu Evlution Comlt? No Tim Comlxity? O(^) Comlxity? O( + N) Dnil. Wl 70 Hill Climing I Alwys oos st il; no ktking Bm s wit uu = Polms? Lol mxim int snt Rnomizing Hill Climing Rnomly isoying uisti Rnom stts ( vy til istiutions ) Pltus Digonl igs Lol Dnil. Wl 7 Dnil. Wl 7

imult Annling Ojtiv: voi lol minim Tniu: Fo t most t us ill liming Wn no imovmnt ossil Coos nom nigo Lt t s in ulity Mov to nigo wit oility - -/T Ru tmtu t (T) ov tim Otiml? I T s slowly noug, will otiml stt Wily us lso WlkAT tm Dnil. Wl 73 3