Daa 6 & 7

UNIT VI
Backtracking: General method, applications-8-queen problem, sum of subsets, graph coloring,

Hamiltonian cycles
Backtracking - General method: Backtracking is a general algorithm for finding all (or some) solutions to
some computational problem that incrementally builds candidates to the solutions, and abandons each
partial candidate c (back tracks!) as soon as it determines that c cannot possibly be completed to a "alid
solution#
Back tracking algorithms try each possibility until they find the right one# $t is a depth-first search of
the set of possible solutions# %uring the search, if an alternati"e does not &ork, the search backtracks to
the choice point, the place &hich presented different alternati"es, and tries the ne't alternati"e# (hen the
alternati"es are e'hausted, the search returns to the pre"ious choice point and tries the ne't alternati"e
there# $f there are no more choice points, the search fails.
Applications: Back tracking method is used to sol"e the follo&ing problems:
8-queen problem, sum of subsets, graph coloring, Hamiltonian cycles
8-queen problem (n-)ueens *roblem):
n queens should be placed in an n'n chess board, in such a &ay that no queen can attack
another queen#
+he follo&ing algorithm computes solution to n-)ueens problem#
,# -irst suppose that e"ery ne& queen &ill be placed in a ne& ro&# +his ensures that no queen can
attack another queen ro&-&ise# +hus n queens can be placed in n different ro& positions#
.# /o&, &e ha"e to calculate the column positions of these n queens# +his can be done in the
follo&ing &ay#
.#, -irst place, first queen in first ro& and first column#
.#. 0alculate the column position of the ne't queen in ne't ro&, in the follo&ing
&ay#
*lace it in the first column of the ne& ro&# $f it collides &ith any of the
pre"iously placed queens (either column-&ise or diagonal-&ise) , then, mo"e it one
step right# $f it goes out of the board, mo"e pre"ious queen one step right, and goto
step .#.#
$f queen can be placed in a particular column in that ro&, &ithout any
collision, then goto step .#.#
$f all n queens are placed in n different column positions, then 1+2*#
2ther&ise goto step .#.#
3 queen collides &ith another queen, column-&ise, if the column positions of the both queens are same#
3 queen collides &ith another queen, diagonal-&ise, if
4%ifference of column positions454%ifference of ro& positions
Ho& to sol"e n )ueens *roblem using backtracking:
,# *lace ith queen in first possible column position in ith ro&#
.# $f you are not able to place ith queen in any column position in ith ro&, then, backtrack and mo"e
(i-,)th queen in (i-,)th ro& one column right#
6# 0ontinue this (steps , and .) until all n queens are placed in n different column positions#
Sum of Subsets:
Gi"en a set 1 of n ob7ects &ith &eights (&,,&.,8,&n) and a positi"e integer 9,&e ha"e to find a subset 1:
of the gi"en set 1, such that, the sum of the elements of subset 1: is equal to 9#
-or e'ample, if a gi"en set 1 5 (,, ., 6, ;) and 9 5 <, then there e'ists sets 1=5 (6, .) and 1!5(,, ;)
&hose sum is equal to 9#
$t can also be noted that some instance of the problem does not ha"e any solution#
-or e'ample, if a gi"en set 8 5 (,, 6, <) and 9 5 >, then no subset occurs for &hich the sum is equal
to 9 5 >#
,

Ho& to sol"e 1um of 1ubsets *roblem using backtracking:
?et (5 (eight added so far to the subset and
+5 +otal remaining &eight in the gi"en set, then,
@"ery node in the partial solution state space tree is represented as ((,i,+)
3 node at le"el i, is promising, iff,
(A&i
m
and
(A+
m
$f at any node, (5m, then the path from root to that node is the solution#
$f a node is not promising, then, &e backtrack from that node to its pre"ious node
Graph Coloring: Graph coloring is a &ay of coloring the "ertices of a graph such that no t&o ad7acent
"ertices share the same color#+he minimum number of colors required to color a graph in this &ay is
called its chromatic number#
Ho& to sol"e graph coloring problem using backtracking::
0hromatic number of a graph can be obtained by using the follo&ing algorithm#
,# $nitialiBe cn to ,# CC cn 5 chromatic number
.# 3pply cn to "erte' ,
6# cc5cnD CC cc5current color
;# &hile(cc can=t be applied to ne't un colored "erte') do
cc5cc-,# CCbacktrack
<# $f(cc
E) F3pply cc to that un colored "erte'D goto 6DG

H# @lse $f(there are any uncolored "ertices) Fcn5cnA,D goto 6DG
># *rint( 0hromatic number 5!, cn)
8# 1top#
/2+@: 3 particular color can=t be applied to a particular "erte', if that "erte' is ad7acent to any of the
"ertices that are already colored &ith that color#
Hamiltonian cycles:
Ho& to sol"e Hamiltonian path problem using backtracking:
*roceed in any possible path, if Hamiltonian cycle is not obtained, then backtrack and take di"ersion#
PROBLEMS
n ueens Problem
,# 1ol"e n queens problem &hen
(i) n5; (ii) n5< (iii) n58
Sum of Subsets
,# ?et &5F,,.,6G and m56# -ind all subsets of the set & that sum to m#
.# ?et &5F>,,,,,6,.;G and m56,# -ind all subsets of the set & that sum to m#
6# ?et &5F<,,E,,.,,6,,<,,8G and m56E# -ind all subsets of the set & that sum to m#
Graph Coloring
, #-ind chromatic number for the follo&ing graphs, and sho& ho& those graphs can be colored#
.
-ig(i): 9ap $ts corresponding graph
-ig(ii) -ig(iii)
Hamiltonian cycles
,# -ind Hamiltonian cycles for the follo&ing graphs#
-ig(i)
-ig(ii)
-ig(iii)
-ig(i")
-ig(")
6
!"#$ %##
Branch and Bound: General method, applications - +ra"elling sales person, EC, knapsack problem- ?0 Branch
and Bound solution, -$-2 Branch and Bound solution
General Metho&:
@-node is the node that is being e'plored currently#
@'ploring a node means generating all its children (branches or li"e-nodes)#
$n Branch and Bound +echnique, all branches(li"e nodes) of an @-node are generated, before one of
the branches(li"e node) becomes an @-node
+here are three types of Branch and Bound techniques# +hey are:
,# ?0BB .#?$-2BB 6#-$-2BB
'pplications: BB can be used to sol"e optimiBation problems such as EC, knapsack problem, tra"eling sales
person problem etc#
$ra(elling sales person problem)$SPP*: +1** can be sol"ed by using BB, in the follo&ing &ay:
,# 2btain reduced matri'(I) for the gi"en cost matri'#
.# 0ompute bound(b) for that reduced matri'#
6# Ioot (@-node5I) consists of all possible paths#
;# ?eft child can be obtained by including a particular edge in the solution set#
<# Iight child can be obtained by e'cluding that particular edge from the solution set#
H# Ieduce left and right childs and compute bounds#
># +he child &ith minimum bound becomes @-node for the ne't iteration#
8# 0ontinue this until either left or right child becomes empty#
J# +hen, solution can be obtained by backtracking from that empty child to root#
9ore e'planations for important steps:
,# -irst, perform ro& reduction
FIo& reduction means subtract e"ery ro& &ith minimum element in that ro&# Io& reduction
sum(rrs) can be obtained by adding all these minimum elements in each ro&#G
+hen, for the resulting matri', perform column reduction
F0olumn reduction means subtract e"ery column &ith minimum element in that column#
0olumn reduction sum(crs) can be obtained by adding all these minimum elements in each column#G
+he resulting matri' is called reduced matri'K
.# Bound b5rrsAcrs

6# ;# $f you include a particular edge(i,7) in the solution set, then,
-delete ith ro& and 7th colunmn#
-put in 7th ro& and ith column
<# $f you e'clude a particular edge(i,7) in the solution set, then,
-put in ith ro& and 7th column#
># /ormally, in e"ery step &e &ill be considering only left child and hence only left child becomes empty
e"entually#
J# (rite do&n all edges &hile you backtrack and find path by connecting them#
-or e"ery E in ILi,7M, compute N(i,7)#
N(i,7)#5min in ro& i A min in col 7 e'cept E in ILi,7M
0onsider that edge(i,7) for &hich you get ma'imum N(i,7)
/ote:
.#
;
+,- .napsac. problem:
Branch and Bound technique is designed for minimiBation problem#
But, EC, knapsack problem is a ma'imiBation problem(as &e ha"e to ma'imiBe profit)# 1o, &e ha"e
to multiply bounds(lo&er and upper) &ith -,#
%o not consider fractions &hile computing upper bound(OB)
0onsider fractions &hile computing lo&er bound(?B)
LCBB)Least Cost Branch an& Boun&*:
+he li"e node &ith minimum(OB-?B) "alue becomes @-node in the ne't iteration#
$f t&o li"e nodes ha"e same (OB-?B) "alue, then the node &ith least OB "alue becomes @-node in
the ne't iteration#
/#/OBB)Last #n /irst Out Branch an& Boun&*:
?et, ?OB5 ?east Opper Bound, 0OB 5 0urrent Opper Bound, P 0?B 5 0urrent ?o&er Bound,
Initialize LUB to infinity.
If (CUB<LUB) then LUB=CUB
If(CLB>LUB) then kill that node
Use Queue to hold all live nodes of E-node.
L#/OBB)/irst in /irst Out Branch an& Boun&*: +his is same as ?$-2BB, e'cept that, here, &e use 1tack to
hold all li"e nodes of @-node#

<

Daa 6 & 7

Uploaded by

Copyright:

Available Formats

Daa 6 & 7

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Daa 6 & 7

Uploaded by

Copyright:

Available Formats

UNIT VI

Backtracking: General method, applications-8-queen problem, sum of subsets, graph coloring,

E) F3pply cc to that un colored "erte'D goto 6DG

You might also like

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.