A Hybrid Genetic Algorithm Approach for OSPF Weight Setting Problem
Transcript of A Hybrid Genetic Algorithm Approach for OSPF Weight Setting Problem
![Page 1: A Hybrid Genetic Algorithm Approach for OSPF Weight Setting Problem](https://reader033.fdocuments.in/reader033/viewer/2022052606/587fd5a61a28ab58248b5661/html5/thumbnails/1.jpg)
Communication Networks E. Mulyana, U. Killat
1
A Hybrid Genetic Algorithm
Approach for OSPF Weight Setting
Problem
PGTS 2002 – Gdansk (Poland) – 23/24.09.2002
![Page 2: A Hybrid Genetic Algorithm Approach for OSPF Weight Setting Problem](https://reader033.fdocuments.in/reader033/viewer/2022052606/587fd5a61a28ab58248b5661/html5/thumbnails/2.jpg)
Communication Networks E. Mulyana, U. Killat
2
Introduction
• OSPF (IGP) use administrative metric
– Not adapt on the traffic situation
Unbalanced load distribution
• Mechanism to increase network utilization and
avoid congestion
– Changing the link weights for a given demand
– The problem is NP-hard
![Page 3: A Hybrid Genetic Algorithm Approach for OSPF Weight Setting Problem](https://reader033.fdocuments.in/reader033/viewer/2022052606/587fd5a61a28ab58248b5661/html5/thumbnails/3.jpg)
Communication Networks E. Mulyana, U. Killat
3
OSPF Routing Problem (1)
• Each link has a cost/weight [1 ... 65535]
• Routers compute paths with Dijkstra‘s
algorithm
• ECMP even-splitting
• Given a demand and a set of weights
Load distribution (does not depend on link
capacities)
![Page 4: A Hybrid Genetic Algorithm Approach for OSPF Weight Setting Problem](https://reader033.fdocuments.in/reader033/viewer/2022052606/587fd5a61a28ab58248b5661/html5/thumbnails/4.jpg)
Communication Networks E. Mulyana, U. Killat
4
OSPF Routing Problem (2)
Find a set
of weights
with minimal
cost
Dijkstra ,
ECMP
Objective (cost)
Function
Network topology
and link capacities
Predicted traffic
demand
Set of weights
Cost value
Utilization (max, av)
![Page 5: A Hybrid Genetic Algorithm Approach for OSPF Weight Setting Problem](https://reader033.fdocuments.in/reader033/viewer/2022052606/587fd5a61a28ab58248b5661/html5/thumbnails/5.jpg)
Communication Networks E. Mulyana, U. Killat
5
Objective Functions
• Objective Function 1 : Staehle, Köhler, Kohlhaas
maximum & average utilization
• Objective Function 2 : Minimizing changes
ij uv ij
uv
ij
t
c
l
Eta
1)(
r
kk
r
kk
k
ww
wwy
,
,
0
1
w1r, w
2r, … , w
kr, … , w
|E|r
w1 , w
2 , … , w
k , … , w
|E|
Ek
kty
Eta
1)(
![Page 6: A Hybrid Genetic Algorithm Approach for OSPF Weight Setting Problem](https://reader033.fdocuments.in/reader033/viewer/2022052606/587fd5a61a28ab58248b5661/html5/thumbnails/6.jpg)
Communication Networks E. Mulyana, U. Killat
6
General Routing Problem
• Lower bound for shortest path (SP) routing
• No SP constraints, no splitting constraints
• LP formulation:
Objective Function
Flow Conservation
Utilization Upper Bound (t)
![Page 7: A Hybrid Genetic Algorithm Approach for OSPF Weight Setting Problem](https://reader033.fdocuments.in/reader033/viewer/2022052606/587fd5a61a28ab58248b5661/html5/thumbnails/7.jpg)
Communication Networks E. Mulyana, U. Killat
7
The Proposed Hybrid-GA
The big picture The population dynamic
Reproduction
Mutation
Heuristic
Search
Best chromosome
Population
50 chromosomes
Selection (parents)
8 chromosomes
Selection
(remove 10%)
Population
45 chromosomes
Offsprings
8 chromosomes
Search result
(1 or 0 chromosome)
Population
53 or 54 chromosomes
Selection
(best 50 chromosomes)
Start
Population
Exit
Condition
Heuristic
Search
Selection
Reproduction
Mutation
Add new
Population
Selection
yes
no
![Page 8: A Hybrid Genetic Algorithm Approach for OSPF Weight Setting Problem](https://reader033.fdocuments.in/reader033/viewer/2022052606/587fd5a61a28ab58248b5661/html5/thumbnails/8.jpg)
Communication Networks E. Mulyana, U. Killat
8
Forming a new generation
• Reproduction
– Crossover
– Arbitrary Mutation
• „Targeted“ Mutation
AV C1 C2 C3 C4
P1 P2
O2 O1
Reproduction
„Targeted“
Mutation
![Page 9: A Hybrid Genetic Algorithm Approach for OSPF Weight Setting Problem](https://reader033.fdocuments.in/reader033/viewer/2022052606/587fd5a61a28ab58248b5661/html5/thumbnails/9.jpg)
Communication Networks E. Mulyana, U. Killat
9
Reproduction
const 2
const 1 0.03
0.53
5 5 6 5 7
1 2 3 3 4 Parent 1 (P1)
Parent 2 (P2)
Intermediate 1
(I1)
Intermediate 2
(I2)
Random 0.81 0.59
5
1
0.02
1
8
0.09
6
3
0.35
5
3 7
4
![Page 10: A Hybrid Genetic Algorithm Approach for OSPF Weight Setting Problem](https://reader033.fdocuments.in/reader033/viewer/2022052606/587fd5a61a28ab58248b5661/html5/thumbnails/10.jpg)
Communication Networks E. Mulyana, U. Killat
10
„Targeted“ Mutation
0.4 1.4 0.1 0.8 0.3 0.6
0.1 0.6 0.7 1.2 0.4 0.6
5
1 6 5
7
1
8 3 3
4
I1
I2
Util. I1
Util. I2
Average
Average
Av - 0.2 Av + 0.2
Utilization
5
1 6 5
7
1
8 3 3
4
3
5 4
7
3
Offspring 1
Offspring 2
0.1
1.4 0.1
1.2
0.3
![Page 11: A Hybrid Genetic Algorithm Approach for OSPF Weight Setting Problem](https://reader033.fdocuments.in/reader033/viewer/2022052606/587fd5a61a28ab58248b5661/html5/thumbnails/11.jpg)
Communication Networks E. Mulyana, U. Killat
11
Heuristic Search
• Individual-based search
• Best chromosome as input
C=A
Improvement?
( fail < treshold )
Apply
Heuristic
B better than C?
C=B
fail = 0 fail ++
yes
Chromosome B
yes no
no
Chromosome C
Chromosome A
![Page 12: A Hybrid Genetic Algorithm Approach for OSPF Weight Setting Problem](https://reader033.fdocuments.in/reader033/viewer/2022052606/587fd5a61a28ab58248b5661/html5/thumbnails/12.jpg)
Communication Networks E. Mulyana, U. Killat
12
Results (1)
• Objective function (2)
• at = 10
Original
(reference) GA
Max. 42.9%
Av. 22.4%
Max. 35.7%
Av. 22.7%
4 weight changes :
(2,1) (3,4) (4,5) (5,6)
![Page 13: A Hybrid Genetic Algorithm Approach for OSPF Weight Setting Problem](https://reader033.fdocuments.in/reader033/viewer/2022052606/587fd5a61a28ab58248b5661/html5/thumbnails/13.jpg)
Communication Networks E. Mulyana, U. Killat
13
A Test Network
![Page 14: A Hybrid Genetic Algorithm Approach for OSPF Weight Setting Problem](https://reader033.fdocuments.in/reader033/viewer/2022052606/587fd5a61a28ab58248b5661/html5/thumbnails/14.jpg)
Communication Networks E. Mulyana, U. Killat
14
Results (2)
![Page 15: A Hybrid Genetic Algorithm Approach for OSPF Weight Setting Problem](https://reader033.fdocuments.in/reader033/viewer/2022052606/587fd5a61a28ab58248b5661/html5/thumbnails/15.jpg)
Communication Networks E. Mulyana, U. Killat
15
Results (3)
![Page 16: A Hybrid Genetic Algorithm Approach for OSPF Weight Setting Problem](https://reader033.fdocuments.in/reader033/viewer/2022052606/587fd5a61a28ab58248b5661/html5/thumbnails/16.jpg)
Communication Networks E. Mulyana, U. Killat
16
Conclusion
• Hybrid genetic algorithm to OSPF routing problem, with „targeted“ mutation and search heuristic
• Propose an objective function to minimize changes
• Compare the result to one with objective function from Staehle, Köhler, Kohlhaas
![Page 17: A Hybrid Genetic Algorithm Approach for OSPF Weight Setting Problem](https://reader033.fdocuments.in/reader033/viewer/2022052606/587fd5a61a28ab58248b5661/html5/thumbnails/17.jpg)
Communication Networks E. Mulyana, U. Killat
17
Thank You !
![Page 18: A Hybrid Genetic Algorithm Approach for OSPF Weight Setting Problem](https://reader033.fdocuments.in/reader033/viewer/2022052606/587fd5a61a28ab58248b5661/html5/thumbnails/18.jpg)
Communication Networks E. Mulyana, U. Killat
18
Convergence
![Page 19: A Hybrid Genetic Algorithm Approach for OSPF Weight Setting Problem](https://reader033.fdocuments.in/reader033/viewer/2022052606/587fd5a61a28ab58248b5661/html5/thumbnails/19.jpg)
Communication Networks E. Mulyana, U. Killat
19
Increasing Traffic