Uninformed Search in Artificail Intelligence

8/9/2019 Uninformed Search in Artificail Intelligence

1/16

UNINFORMED SEARCHUNINFORMED SEARCH

-BFS-BFS

-DFS-DFS

-DFIS-DFIS

- Bidirectional- Bidirectional

INT404K3305


2/16

•State-space search is the process of searching through a state space fora solution by making explicit a sufficient portion of an implicit state-

space graph to include a goal node.

– Hence, initially V={!, "here is the start node#

– "hen is expanded, its successors are generated and those nodes areadded to V and the associated arcs are added to $.

– %his process continues until a goal node is generated &included in V' andidentified &by goal test'

•(uring search, a node can be in one of the three categories)

– *ot generated yet &has not been made explicit yet' – OPEN) generated but not expanded

– CLOSED) expanded

– earch strategies differ mainly on ho" to select an +$* node for

expansion at each step of search


3/16

A General State-Space Search Algorithm

open )= {!# closed )={!#repeat

n )= select &open'# select one node from open for expansion

if n is a goal

then exit "ith success# delayed goal testing

expand &n'

generate all children of n

put these ne"ly generated nodes in open &check duplicates'

put n in closed &check duplicates'

until open = {!#

exit "ith failure


4/16

Some Issues•earch process constructs a search tree, "here

– root is the initial state , and – leaf nodes are nodes

•not yet been expanded &i.e., they are in +$* list' or

•ha/ing no successors &i.e., they0re 1deadends1'

•ome important issue that arises

•%he direction in "hich conduct the search&for"ard /s. back"ard

reasoning'

•Ho" to select applicable rules&matching'.

•Ho" to represent each node of search process&the kno"ledge

representation problem'

•earch tree /s. search graph


5/16


6/16

Algorithm ( )readth-*irst Search

+,

1. Create a variable called NODE-LIST and set it to the initial state.

2. Until a goal state is found or NODE-LIST is e!t"#

$a%&eove the first eleent fro NODE-LIST and call it E. If NODE-LIST

'as e!t"( )uit.

$b% *or each 'a" that each rule can atch the state described in E do#

(i) Apply the rule to generate a new state,

(ii) If the new state is a goal state, quit and return this state.

(iii) Otherwise, add the new state to the end of NODE-LI!.


7/16

!o Leels of a )readth-*irst

Search !ree

+.


8/16

+readth,*irst

- co!lete search tree of

de!th d 'here each non,leafnode has b children( has a

total of 1 / b / b2 / ... / bd

$b$d/1% , 1%$b,1% nodes

- Tie co!leit" $ of nodes

generated%# $bd% - 6!ace co!leit" $aiu

length of 78N%# $bd%

s/

#

#0+

#0dd

+

/

- *or a co!lete search tree of de!th 12( 'here ever" node atde!ths 0( ...( 11 has 10 children and ever" node at de!th 12 has 0

children( there are 1 / 10 / 100 / 1000 / ... / 1012 $1013 ,1%9 $1012% nodes in the co!lete search tree.

: +*6 is suitable for !robles 'ith shallo' solutions


9/16

Algorithm ( Depth-*irst Search

+1

1. If the initial state is a goal state( )uit and return success.

2. ther'ise( do the follo'ing until success or failure is signaled#

(a) "enerate a su##essor, E, of the initial state. If there are no $ore

su##essors, signal failure.

(%) &all Depth-'irst ear#h with E as the initial state.

(#) If su##ess is returned, signal su##ess. Otherwise #ontinue in this loop.


10/16

A Depth-*irst Search !ree

+2


11/16

;e!th,*irst $;*6%: lgorith outline#

- l'a"s select fro the 78N the node 'ith thegreatest de!th for e!ansion( and !ut all ne'l"

generated nodes into 78N

- 78N is organience( onl" does =chronological bac@trac@ing=

goal


12/16

;e!th,*irst Iterative ;ee!ening $;*I;%

: +* and ;* both have e!onential tie co!leit" $bd%

+* is #o$plete but has e!onential s!ace co!leit";* has linear spa#e #o$pleity but is inco!lete

: 6!ace is often a harder resource constraint than tie

: Can 'e have an algorith that - Is co!lete

- >as linear s!ace co!leit"( and - >as tie co!leit" of $bd%

: ;*I; b" Korf in 19B5 $1 "ears after D%

*irst do ;*6 to de!th 0 $i.e.( treat start node as

having no successors%( then( if no solution found(do ;*6 to de!th 1( etc.

until solution found do

DFS with depth bound d

d = d+1


13/16

;e!th,*irst Iterative ;ee!ening $;*I;%

: &o$plete $iterativel" generate all nodes u! to de!th

d%

: Opti$al+Ad$issi%le if all o!erators have the sae

cost. ther'ise( not o!tial but does guarantee

finding solution of shortest length $li@e +*%.

: Linear spa#e #o$pleity $bd%( $li@e ;*%

: !i$e #o$pleity is a little 'orse than +*6 or ;*6

because nodes near the to! of the search tree are

generated ulti!le ties( but because alost all of

the nodes are near the botto of a tree( the 'orst

case tie co!leit" is still e!onential( $bd%


14/16

;e!th,*irst Iterative ;ee!ening

: If branching factor is b and solution is at de!th d( then nodes at

de!th d are generated once( nodes at de!th d,1 are generated

t'ice( etc.( and node at de!th 1 is generated d ties.

>ence

total$d% bd / 2b$d,1% / ... / db


15/16


16/16

Uninformed Search in Artificail Intelligence

Documents

Transcript of Uninformed Search in Artificail Intelligence