Generalized partition crossover for the traveling salesman problem

The Traveling Salesman Problem (TSP) is a well-studied combinatorial optimization problem with a wide spectrum of applications and theoretical value. We have designed a new recombination operator known as Generalized Partition Crossover (GPX) for the TSP. GPX is unique among other recombination operators for the TSP in that recombining two local optima produces new local optima with a high probability. Thus the operator can 'tunnel' between local optima without the need for intermediary solutions. The operator is respectful, meaning that any edges common between the two parent solutions are present in the offspring, and transmits alleles, meaning that offspring are comprised only of edges found in the parent solutions. We design a hybrid genetic algorithm, which uses local search in addition to recombination and selection, specifically for GPX. We show that this algorithm outperforms Chained Lin-Kernighan, a state-of-the-art approximation algorithm for the TSP. We next analyze these algorithms to determine why the algorithms are not capable of consistently finding a globally optimal solution. Our results reveal a search space structure which we call 'funnels' because they are analogous to the funnels found in continuous optimization. Funnels are clusters of tours in the search space that are separated from one another by a non-trivial distance. We find that funnels can trap Chained Lin-Kernighan, preventing the search from finding an optimal solution. Our data indicate that, under certain conditions, GPX can tunnel between funnels, explaining the higher frequency of optimal solutions produced by our hybrid genetic algorithm using GPX.

Name thesis:degree_name

Master of Science (M.S.)

Level thesis:degree_level

Masters

Discipline thesis:degree_discipline

Computer Science

Grantor dc:publisher

Colorado State University. Libraries

Year dc:date.issued

2011

Authors dc:creator

• Hains, Douglas R., author
• Whitley, L. Darrell, advisor
• Howe, Adele E., committee member
• Mueller, Jennifer L., committee member

• genetic algorithms
• Traveling Salesman Problem
• search space
• local search

Statement dc:rights

• Copyright and other restrictions may apply. User is responsible for compliance with all applicable laws. For information about copyright law, please see https://libguides.colostate.edu/copyright.

Language dc:language.iso

eng, English

Identifier URI

https://doi.org/10.25675/3.018670

OAI identifier oai:identifier

oai:mountainscholar.org:10217/47314