1
Greedy Randomized Adaptive Search and Variable Neighbourhood Search for the minimu
m labelling spanning tree problem
Kuo-Hsien Chuang 2008/11/05
2
Introduction
Output graphFitness = 2
Input graph
3
Literature review• Maximum vertex covering algorithm
4
Literature review
• MVCA applying Pilot method– Let C = empty set of labels– Set C = {all c 屬於 (L – C), min(comp(C + c))}
5
Exploited metaheuristics
• MGA
6
Exploited metaheuristics• MGA
7
Exploited metaheuristics
• MGA
8
Exploited metaheuristics• GRASP
9
Exploited metaheuristics
10
Exploited metaheuristics
11
Exploited metaheuristics
12
Exploited metaheuristics• VNS
13
Exploited metaheuristics
14
Exploited metaheuristics
15
Exploited metaheuristics
16
Computational results
17
18
19
20
Conclusion
• All the results allow us to state that VNS and GRASP are fast and extremely effective metaheuristics for the MLST problem
• Future research :
an algorithm based on Ant Colony Optimisation (ACO) is currently under study in order to try to obtain a larger diversification capability by extending the current greedy MVCA local search.