2006
DOI: 10.1016/j.chaos.2005.08.114
|View full text |Cite
|
Sign up to set email alerts
|

An new self-organizing maps strategy for solving the traveling salesman problem

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1
1
1

Citation Types

2
32
0

Year Published

2006
2006
2024
2024

Publication Types

Select...
6
1

Relationship

0
7

Authors

Journals

citations
Cited by 40 publications
(34 citation statements)
references
References 10 publications
2
32
0
Order By: Relevance
“…As a consequence, the time required to execute one iteration, on an average, becomes almost equal for both FRAG GA and OX SIM. Similar observations are also made when the proposed method is compared with other GAs [1,6] and other methods like Self Organizing Map (SOM) [10].…”
Section: Comparison With Other Ga Approaches For Tspsupporting
confidence: 63%
See 3 more Smart Citations
“…As a consequence, the time required to execute one iteration, on an average, becomes almost equal for both FRAG GA and OX SIM. Similar observations are also made when the proposed method is compared with other GAs [1,6] and other methods like Self Organizing Map (SOM) [10].…”
Section: Comparison With Other Ga Approaches For Tspsupporting
confidence: 63%
“…FRAG GA is implemented in C on Pentium-4 (1.2 GHz) and the results are compared with those obtained using SWAP GATSP [6], MMGA [4], IGA [5], OX SIM, (standard GA with order crossover and simple inversion mutation) [1] MOC SIM (Modified order crossover and SIM), and self organizing map (SOM) [10] for solving TSP. For fair comparison with the above mentioned methods Lin-Kernighan (LK) heuristic is not used with FRAG GA, whereas, for comparison with HeSEA [16] and other LK based methods each chromosome in FRAG GALK is updated probabilistically with 20 runs of consecutive chained LK and mutation (as recommended in [16]).…”
Section: Resultsmentioning
confidence: 99%
See 2 more Smart Citations
“…4 shows that from the 36 problems tested, 14 have the worst results with higher costs with the Soft WTA technique than with the Hard WTA: rd100, a280, ch130, bier127, kroA100, kroC100, kroE100, brazil58, pr107, eil51, gil262, lin318, fl417 and rat575. (Jin et al, 2003); • the Co-Adaptive Network (CAN), which uses the idea of cooperation among neighboring neurons and uses a number of neurons that is higher than the number of cities in the problem (Cochrane & Beasley, 2003); • the Real-Valued Antibody Network (RABNET), which uses a mechanism to stabilize the winning neurons and the centroids of the groups of cities for growth and pruning of the network (Massutti & Castro, 2009); • the Modified Growing Ring Self-Organizing Network (MGSOM) incorporates other initialization methods for the weights in the network, with other adaptation parameters proposed for the SOM network and other indexing forms for the order of cities (Bai et al, 2006); • the MSOM, which consists in a hybrid technique with Self-Organizing Maps (SOM) and evolutionary algorithms to solve the TSP, called Memetic Neural Network (Créput & Kouka, 2007); and • the technique of building a decision tree with minimum amplitude to choose the candidate cities for path exchange with the Lin-Kernighan of 2 up to 5-opt techniques (Kelsgaun, 2000). The comparisons are shown in Table 3, where 16 of the 24 problems tested have better results with the technique proposed using the 2-opt route improving technique.…”
Section: Resultsmentioning
confidence: 99%