ART/SOFM: A Hybrid Approach to the TSP

Narayan Vishwanathan; Donald C. Wunsch

doi:10.1109/IJCNN.2001.938771

ART/SOFM: A Hybrid Approach to the TSP

Narayan Vishwanathan
, Donald C. Wunsch

Research output: Contribution to journal › Article › peer-review

Abstract

We present a new method of solving large scale travelling salesman problem (TSP) instances using a combination of adaptive resonance theory (ART) and self organizing feature maps (SOFM). We divide our algorithm into three phases: phase one uses ART to form clusters of cities; phase two uses a novel modification of the traditional SOFM algorithm to solve a slight variant of the TSP in each cluster of cities; and phase three uses another version of the SOFM to link all the clusters. The experimental results show that our algorithm finds approximate solutions which are about 13% longer than those reported by the chained Lin Kernighan method for problem sizes of 14,000 cities

Original language	American English
Journal	Proceedings of the International Joint Conference on Neural Networks, 2001. IJCNN '01
DOIs	https://doi.org/10.1109/IJCNN.2001.938771
State	Published - Jan 1 2001

Keywords

ART Neural Nets
ART Neural Network
Adaptive Resonance Theory
Approximate Solutions
Approximation Theory
Mathematics Computing
Self Organizing Feature Maps
Self-Organising Feature Maps
Travelling Salesman Problems

Disciplines

Electrical and Computer Engineering

Access to Document

10.1109/IJCNN.2001.938771License: Unspecified

ART SOFM A Hybrid Approach to the TSPFinal published version, 421 KBLicense: Unspecified

Cite this

@article{35f00bfa71d34a68ab4b3f6d7a28c17b,

title = "ART/SOFM: A Hybrid Approach to the TSP",

abstract = " We present a new method of solving large scale travelling salesman problem (TSP) instances using a combination of adaptive resonance theory (ART) and self organizing feature maps (SOFM). We divide our algorithm into three phases: phase one uses ART to form clusters of cities; phase two uses a novel modification of the traditional SOFM algorithm to solve a slight variant of the TSP in each cluster of cities; and phase three uses another version of the SOFM to link all the clusters. The experimental results show that our algorithm finds approximate solutions which are about 13\% longer than those reported by the chained Lin Kernighan method for problem sizes of 14,000 cities",

keywords = "ART Neural Nets, ART Neural Network, Adaptive Resonance Theory, Approximate Solutions, Approximation Theory, Mathematics Computing, Self Organizing Feature Maps, Self-Organising Feature Maps, Travelling Salesman Problems",

author = "Narayan Vishwanathan and Wunsch, \{Donald C.\}",

year = "2001",

month = jan,

day = "1",

doi = "10.1109/IJCNN.2001.938771",

language = "American English",

journal = "Proceedings of the International Joint Conference on Neural Networks, 2001. IJCNN '01",

}

TY - JOUR

T1 - ART/SOFM: A Hybrid Approach to the TSP

AU - Vishwanathan, Narayan

AU - Wunsch, Donald C.

PY - 2001/1/1

Y1 - 2001/1/1

N2 - We present a new method of solving large scale travelling salesman problem (TSP) instances using a combination of adaptive resonance theory (ART) and self organizing feature maps (SOFM). We divide our algorithm into three phases: phase one uses ART to form clusters of cities; phase two uses a novel modification of the traditional SOFM algorithm to solve a slight variant of the TSP in each cluster of cities; and phase three uses another version of the SOFM to link all the clusters. The experimental results show that our algorithm finds approximate solutions which are about 13% longer than those reported by the chained Lin Kernighan method for problem sizes of 14,000 cities

AB - We present a new method of solving large scale travelling salesman problem (TSP) instances using a combination of adaptive resonance theory (ART) and self organizing feature maps (SOFM). We divide our algorithm into three phases: phase one uses ART to form clusters of cities; phase two uses a novel modification of the traditional SOFM algorithm to solve a slight variant of the TSP in each cluster of cities; and phase three uses another version of the SOFM to link all the clusters. The experimental results show that our algorithm finds approximate solutions which are about 13% longer than those reported by the chained Lin Kernighan method for problem sizes of 14,000 cities

KW - ART Neural Nets

KW - ART Neural Network

KW - Adaptive Resonance Theory

KW - Approximate Solutions

KW - Approximation Theory

KW - Mathematics Computing

KW - Self Organizing Feature Maps

KW - Self-Organising Feature Maps

KW - Travelling Salesman Problems

UR - https://scholarsmine.mst.edu/ele_comeng_facwork/1461

U2 - 10.1109/IJCNN.2001.938771

DO - 10.1109/IJCNN.2001.938771

M3 - Article

JO - Proceedings of the International Joint Conference on Neural Networks, 2001. IJCNN '01

JF - Proceedings of the International Joint Conference on Neural Networks, 2001. IJCNN '01

ER -

ART/SOFM: A Hybrid Approach to the TSP

Abstract

Keywords

Disciplines

Access to Document

Other files and links

Fingerprint

Cite this