Final Year Presentation in EWU
-
Upload
robiul-islam -
Category
Engineering
-
view
54 -
download
2
Transcript of Final Year Presentation in EWU
EVOLUTIONARY ALGORITHM FOR
GRAPH COLORING PROBLEM
Presented by
Robiul Islam
2009-2-60-004
And
Arup Kumar Pramanik
2009-2-60-008
East West University
Date : 28 April, 2013
1
SUPERVISOR
Professor Dr. Mozammel Huq Azad Khan
Dept. of Computer Science & Engineering
East West University
2
KEY TERM
Evolutionary Algorithm
Binary Encoding
Mutation
Adding new population
Deterministic process
3
GRAPH COLORING PROBLEM
A coloring of simple graph is the assignment of a
color to each vertex of the graph so that no two
adjacent vertices are assigned the same color. The
chromatic number of a graph is the least number of
colors needed for a coloring of this graph.
* Well-Known NP-hard Problem
* Two adjacency nodes does not contain
same colour
* Uses minimum number of colours
* Also known as vertex colouring problem
4
EVOLUTIONARY ALGORITHM
Evolutionary algorithms (EA) are search
algorithm based on the mechanics of natural
selection and natural genetics
In every generation, a new set of artificial
creatures (chromosome) is created using bits and
pieces of the fittest of the old, an occasional new
part is tried for good measure.
5
DETAILS OF MYCIEL3.COL GRAPH
Node = 11
Edge = 20
Initialization Color = Maximum Out degree +1, which is upper bound of chromatic number
Represent binary matrix with Initialization Color*Node
Here number of row = 6
Number of column = 11
Row represent Color
Column represent Node
6
ENCODING TECHNIQUE
0 0 1 0 0 1 0 0 0 0 1
0 0 0 0 0 0 1 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 1 0 0 1 0 0 0
1 0 0 0 0 0 0 0 0 0 0
0 1 0 1 0 0 0 0 1 1 0
Vertices
1
2
3
4
5
6
Colors
1 2 3 4 5 6 7 8 9 10 11
7
ALGORITHM DESCRIPTION : FITNESS
0 0 1 0 0 1 0 0 0 0 1
0 0 0 0 0 0 1 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 1 0 0 1 0 0 0
1 0 0 0 0 0 0 0 0 0 0
0 1 0 1 0 0 0 0 1 1 0
Vertices
1
2
3
4
5
6
Colors
1 2 3 4 5 6 7 8 9 10 11
Invalid Color
Valid Color
Unused Color
8
ALGORITHM DESCRIPTION : FITNESS
(CONTINUE)
0 0 1 0 0 1 0 0 0 0 1
0 0 0 0 0 0 1 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 1 0 0 1 0 0 0
1 0 0 0 0 0 0 0 0 0 0
0 1 0 1 0 0 0 0 1 1 0
Invalid Row : 3
Valid Row : 2
Unused Row :1
Fitness = Invalid Row * (maximum out degree+1) + Valid Row
Here myciel3.col data file
Fitness Value = 3*6+2
= 20
1 2 3 4 5 6 7 8 9 10 11
1
2
3
4
5
6
Two edges share same ages those
colors are not valid
9
ALGORITHM DESCRIPTION : CORRECTION
0 0 1 0 0 0 0 0 0 0 1
0 0 0 0 0 0 1 0 0 0 0
0 0 0 0 0 1 0 0 0 0 0
0 0 0 0 1 0 0 1 0 0 0
1 0 0 0 0 0 0 0 0 0 0
0 1 0 1 0 0 0 0 1 1 0
1 2 3 4 5 6 7 8 9 10 11
1
2
3
4
5
6
Invalid Row : 2
Valid Row : 4
Fitness Value = 2*6+4
= 16 10
ALGORITHM DESCRIPTION : CORRECTION
(CONTINUE )
0 0 1 0 0 0 0 0 0 0 1
0 0 0 0 0 0 1 0 0 0 0
0 0 0 0 1 1 0 0 0 0 0
0 0 0 0 0 0 0 1 0 0 0
1 0 0 0 0 0 0 0 0 0 0
0 1 0 1 0 0 0 0 1 1 0
Invalid Row : 1
Valid Row : 5
Fitness Value = 1*6+5
= 11
1 2 3 4 5 6 7 8 9 10 11
1
2
3
4
5
6
11
ALGORITHM DESCRIPTION : CORRECTION
(CONTINUE)
0 0 1 0 0 0 0 0 0 0 1
0 0 0 0 0 0 1 0 0 0 0
0 0 0 0 1 1 0 0 0 0 0
0 0 0 1 0 0 0 1 0 0 0
1 0 0 0 0 0 0 0 0 0 0
0 1 0 0 0 0 0 0 1 1 0
Invalid Row : 0
Valid Row : 6
Fitness Value = 0*6+6
= 6
1 2 3 4 5 6 7 8 9 10 11
1
2
3
4
5
6
12
ALGORITHM DESCRIPTION
Remove duplicate chromosome
Copying total number of population to a
temporary population
13
ALGORITHM DESCRIPTION : MUTATION
0 0 1 0 0 0 0 0 0 0 1
0 0 0 0 0 0 1 0 0 0 0
0 0 0 0 1 1 0 0 0 0 0
0 0 0 1 0 0 0 1 0 0 0
1 0 0 0 0 0 0 0 0 0 0
0 1 0 0 0 0 0 0 1 1 0
1 2 3 4 5 6 7 8 9 10 11
1
2
3
4
5
6
0 0 1 0 0 0 0 0 0 0 1
0 0 0 0 0 0 1 0 0 0 1
0 0 0 0 1 1 0 0 0 0 0
0 0 0 1 0 0 0 0 0 0 0
1 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 1 1 0
1 2 3 4 5 6 7 8 9 10 11
1
2
3
4
5
6
Basic Chromosome Mutated Chromosome
Randomly select one bit in a row with low probability If it’s zero convert into one If it’s one convert into zero 14
ALGORITHM DESCRIPTION : REPAPERING
POPULATION
0 0 1 0 0 0 0 0 0 0 1
0 0 0 0 0 0 1 0 0 0 1
0 0 0 0 1 1 0 0 0 0 0
0 0 0 1 0 0 0 0 0 0 0
1 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 1 1 0
1 2 3 4 5 6 7 8 9 10 11
1
2
3
4
5
6
Mutated Chromosome
0 0 1 0 0 0 0 0 0 0 0
0 0 0 0 0 0 1 0 0 0 1
0 1 0 0 1 1 0 0 0 0 0
0 0 0 1 0 0 0 0 0 0 0
1 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 1 1 1 0
1
2
3
4
5
6
1 2 3 4 5 6 7 8 9 10 11
After Reaper Chromosome
15
ALGORITHM DESCRIPTION
Calculate fitness of mutated population
Corrected mutation population
Remove duplicate chromosome of mutated
population
16
MERGING TEMPORARY POPULATION TO
INITIAL POPULATION ACCORDING TO MINIMUM
FITNESS VALUE AND ADDING
Adding new population of initial population replace
worst population
6 5 6 5 6 5
6 4 4 5 6 5
Fitness Value of
Population
Fitness Value of
Mutated Population
4 5 4 5 5 5Fitness Value after
Merge
17
DETERMINISTIC PROCESS
Before After
Valid Color 6 Valid Color 4
18
EXPERIMENTAL RESULT
Data File (𝐺) Node Edge 𝜒(𝐺) EAGCP [1] [2] [3]
myciel3.col 11 20 4 4 4 4 4
myciel4.col 23 71 5 5 5 5 5
queen5_5.col 25 160 5 5 5 5 5
queen6_6.col 36 290 7 8 7 7 7
myciel5.col 47 236 6 6 6 6 6
huck.col 74 301 11 11 11 11 11
jean.col 80 254 10 10 10 10 10
anna.col 138 493 11 12 11 11 11
david.col 87 406 11 12 11 11 11
19
EXPERIMENTAL RESULT
Data File :myciel3.col
Node: 11
Edge: 20
Generation: 51
Chi (G): 4
EAGCP: 4
Population Size: 50
Maximum Color: 6
Mutation Probability: 10%
Additional Probability: 10%20
EXPERIMENTAL RESULT (CONTINUE)
Date File : myciel4.col
Node: 23
Edge: 71
Generation: 163
Chi (G): 5
EAGCP: 5
Population Size: 50
Maximum Color: 12
Mutation Probability: 10%
Additional Probability: 10%21
EXPERIMENTAL RESULT (CONTINUE)
Date File : myciel5.col
Node: 47
Edge: 236
Generation: 765
Chi (G): 6
EAGCP: 6
Population Size: 150
Maximum Color: 24
Mutation Probability: 10%
Additional Probability: 10%22
EXPERIMENTAL RESULT (CONTINUE)
Data File: queen5_5.col
Node: 25
Edge: 160
Generation: 845
Chi (G): 5
EAGCP: 5
Population Size: 200
Maximum Color: 17
Mutation Probability: 15%
Additional Probability: 10%23
EXPERIMENTAL RESULT (CONTINUE)
Date File : huck.col
Node: 71
Edge: 301
Generation: 1897
Chi (G): 11
EAGCP: 11
Population Size: 500
Maximum Color: 54
Mutation Probability: 10%
Additional Probability: 10%24
CONCLUSION
In this thesis work we have focused on a better
minimize chromatic number with proper EA step
for GCP
It helps mutation, evaluate, immune system and
also reduce colour dynamically
Our best result of large dataset is huck.col which
has 71 node and 301 edge and find expected color
in this graph
25
FUTURE WORK
Used our algorithm in different stranded data set
Optimal result of large data set
Reducing the time complexity
26
THANK YOU
27