Yiyuan Gong

Parallel genetic algorithms on line topology of heterogeneous computing resources

This paper evaluates a parallel genetic algorithm (GA) on the line topology of heterogeneous computing resources. Evolution process of parallel GAs is investigated on two types of arrangements of heterogeneous computing resources: the ascending and descending order arrangement of computing capability. Their differences in chromosome variety, migration frequency and solution quality are investigated. The results in this paper can help to design parallel GAs in grid computing environments.

Migration Effects of Parallel Genetic Algorithms on Line Topologies of Heterogeneous Computing Resources

This paper investigates migration effects of parallel genetic algorithms (GAs) on the line topology of heterogeneous computing resources. Evolution process of parallel GAs is evaluated experimentally on two types of arrangements of heterogeneous computing resources: the ascending and descending order arrangements. Migration effects are evaluated from the viewpoints of scalability, chromosome diversity, migration frequency and solution quality. The results reveal that the performance of parallel GAs strongly depends on the design of the chromosome migration in which we need to consider the arrangement of heterogeneous computing resources, the migration frequency and so on. The results contribute to provide referential scheme of implementation of parallel GAs on heterogeneous computing resources.

Iterative Parallel Genetic Algorithms Based on Biased Initial Population

This paper proposes an iterative parallel genetic algorithm with biased initial population to solve large-scale combinatorial optimization problems. The proposed scheme employs a master-slave collaboration in which the master node manages searched space of slave nodes and assigns seeds to generate initial population to slaves for their restarting of evolution process. Our approach allows us as widely as possible to search by all the slave nodes in the beginning period of the searching and then focused searching by multiple slaves on a certain spaces that seems to include good quality solutions. Computer experiment shows the effectiveness of our proposed scheme.

Iterative parallel and distributed genetic algorithms with biased initial population

This work proposes an iterative parallel and distributed genetic algorithm with biased initial population to solve large-scale combinatorial optimization problems. The proposed scheme is a master-slave style in which a master node manages searched space of slave nodes and assigns seeds to generate initial population to slaves for their restarting of evolution process. Our approach allows us as wide as possible searching by all the slave nodes in the beginning periods of the searching and then focused searching by multiple slaves on a certain spaces that seems to include good quality solutions. Computer experiment shows the effectiveness of our proposed scheme.

Experimental evaluation of a parallel genetic algorithm on line topology of heterogeneous computing resources

The paper investigates how the arrangement of heterogeneous computing resources affects evolution progress in a parallel genetic algorithm. This analysis is based on two types of computing groups, one sorted in ascending order and another one in descending order of computing capability. The results can help to design an efficient parallel genetic algorithm in grid computing environments.