pengenalan distributed artificial intelligence -...
TRANSCRIPT
Swarm Intelligence
12/9/2009 1LSR, Sistem Cerdas: IK461
Swarm Intelligence
• Swarm Intelligence is a new computational and behavioral metaphor for solving distributed problems;
• It is based on the principles underlying the behavior of natural systems consisting of many agents, such as ant colonies and bird flocks .
• The approach emphasizes distributedness, direct or indirect interactions among relatively simple agents, flexibility, and robustness
12/9/2009 2LSR, Sistem Cerdas: IK461
Food sourceNest
Ant Foraging
12/9/2009 3LSR, Sistem Cerdas: IK461
DAI Applications
• Swarm Intelligence– Multi (collaborative) robots– Routing problems– Communication networks– Distributed control– Various forms of optimizations– Etc.
• Software Agents– Distributed data base– Computer networks– Security– Finance, Banks– E-commerce– Etc.
12/9/2009 4LSR, Sistem Cerdas: IK461
Text Books
• Swarm Intelligence : From Natural to Artificial Systems, Eric Bonabeau et al., Oxford Univ. Press, 1999
• Ant Colony Optimization, M.Dorigo, MIT Press, 2004
• Intelligent Software Agents, Richard Murch & Tony Johnson, Prentice-Hall, 1999
• Artificial Intelligence : A Modern Approach, Russel & Norvig, …., 2003
12/9/2009 5LSR, Sistem Cerdas: IK461
12/9/2009 6LSR, Sistem Cerdas: IK461
12/9/2009 7LSR, Sistem Cerdas: IK461
12/9/2009 8LSR, Sistem Cerdas: IK461
12/9/2009 9LSR, Sistem Cerdas: IK461
12/9/2009 10LSR, Sistem Cerdas: IK461
12/9/2009 11LSR, Sistem Cerdas: IK461
12/9/2009 12LSR, Sistem Cerdas: IK461
12/9/2009 13LSR, Sistem Cerdas: IK461
12/9/2009 14LSR, Sistem Cerdas: IK461
12/9/2009 15LSR, Sistem Cerdas: IK461
12/9/2009 16LSR, Sistem Cerdas: IK461
12/9/2009 17LSR, Sistem Cerdas: IK461
12/9/2009 18LSR, Sistem Cerdas: IK461
12/9/2009 19LSR, Sistem Cerdas: IK461
12/9/2009 20LSR, Sistem Cerdas: IK461
12/9/2009 21LSR, Sistem Cerdas: IK461
12/9/2009 22LSR, Sistem Cerdas: IK461
12/9/2009 23LSR, Sistem Cerdas: IK461
12/9/2009 24LSR, Sistem Cerdas: IK461
12/9/2009 25LSR, Sistem Cerdas: IK461
12/9/2009 26LSR, Sistem Cerdas: IK461
12/9/2009 27LSR, Sistem Cerdas: IK461
12/9/2009 28LSR, Sistem Cerdas: IK461
12/9/2009 29LSR, Sistem Cerdas: IK461
12/9/2009 30LSR, Sistem Cerdas: IK461
12/9/2009 31LSR, Sistem Cerdas: IK461
12/9/2009 32LSR, Sistem Cerdas: IK461
12/9/2009 33LSR, Sistem Cerdas: IK461
12/9/2009 34LSR, Sistem Cerdas: IK461
12/9/2009 35LSR, Sistem Cerdas: IK461
12/9/2009 36LSR, Sistem Cerdas: IK461
12/9/2009 37LSR, Sistem Cerdas: IK461
12/9/2009 38LSR, Sistem Cerdas: IK461
12/9/2009 39LSR, Sistem Cerdas: IK461
12/9/2009 40LSR, Sistem Cerdas: IK461
12/9/2009 41LSR, Sistem Cerdas: IK461
12/9/2009 42LSR, Sistem Cerdas: IK461
12/9/2009 43LSR, Sistem Cerdas: IK461
12/9/2009 44LSR, Sistem Cerdas: IK461
Flow Chart Ant System Untuk TSP
12/9/2009 45LSR, Sistem Cerdas: IK461
12/9/2009 46LSR, Sistem Cerdas: IK461
12/9/2009 47LSR, Sistem Cerdas: IK461
Kontruksi Ant System
1. Kontruksi Graph sebagai representasi masalah
2. Fungsi heuristik (η).
3. Probabilistik Transition Rule.
4. Pheromone update (τ).
5. Metode pemberhentian.
12/9/2009 48LSR, Sistem Cerdas: IK461
Jarak antar kota i dan j:
Nilai heuristik (η):
2 2 1/ 2[( ) ( ) ]ij i j i jd x x y y
1ij
ijd
12/9/2009 49LSR, Sistem Cerdas: IK461
Transition rule:
Transition rule untuk TSP:Dimana:
α dan β = parameter yang mengontrol bobot relatif.
τij(t) = pheromone trail pada edge yang menghubungkan city i ke j.
( )( )
( )ki
ij ijk
ij
ij ijt j
tP t
t
( ) 1/( )
( ) 1/ki
ij ijk
ij
ij ijt j
t dP t
t d
12/9/2009 50LSR, Sistem Cerdas: IK461
Formula:
( ) (1 ). ( ) . ( ) dim
( )
/ ( ), ( , ) ( )( )
0,
/ , ( , )( )
0,
( 1) ( )
e
ij ij ij
k
ij ij
k k
k
ij
k
ij
ij ij
t t e t ana
t
Q L t jika i j T tt
sebaliknya
dan
Q L jika i j Tt
sebaliknya
t t
Catatan:
ρ = evaporation ([0..1]).
k = semut ke k.
t = iterasi ke t.
T+ = the best solution.
Tk = solusi terbaik pd suatu
iterasi 12/9/2009 51LSR, Sistem Cerdas: IK461
12/9/2009 52LSR, Sistem Cerdas: IK461
12/9/2009 53LSR, Sistem Cerdas: IK461
12/9/2009 54LSR, Sistem Cerdas: IK461