A comment on bio-inspired optimisation via GPU architecture: the genetic algorithm workload

Authors:
Paula Prata;Paulo Fazendeiro;Pedro Sequeira;Chandrashekhar Padole
Affiliations:
Department of Informatics, Instituto de Telecomunicações (IT) Portugal, University of Beira Interior, Portugal;Department of Informatics, Instituto de Telecomunicações (IT) Portugal, University of Beira Interior, Portugal;Department of Informatics, Instituto de Telecomunicações (IT) Portugal, University of Beira Interior, Portugal;Department of Informatics, Instituto de Telecomunicações (IT) Portugal, University of Beira Interior, Portugal
Venue:
SEMCCO'12 Proceedings of the Third international conference on Swarm, Evolutionary, and Memetic Computing
Year:
2012

Citing 8
Cited 0

Genetic Algorithms in Search, Optimization and Machine Learning

Genetic Algorithms in Search, Optimization and Machine Learning
Handbook of Evolutionary Computation

Handbook of Evolutionary Computation
Multi-Objective Optimization Using Evolutionary Algorithms

Multi-Objective Optimization Using Evolutionary Algorithms
Evolutionary Algorithms for Solving Multi-Objective Problems (Genetic and Evolutionary Computation)

Evolutionary Algorithms for Solving Multi-Objective Problems (Genetic and Evolutionary Computation)
Parallelization of an evolutionary algorithm on a platform with multi-core processors

EA'09 Proceedings of the 9th international conference on Artificial evolution
Parallel genetic algorithm on the CUDA architecture

EvoApplicatons'10 Proceedings of the 2010 international conference on Applications of Evolutionary Computation - Volume Part I
Towards cost-effective bio-inspired optimization: a prospective study on the GPU architecture

SEMCCO'11 Proceedings of the Second international conference on Swarm, Evolutionary, and Memetic Computing - Volume Part II
Semantic constraints for membership function optimization

IEEE Transactions on Systems, Man, and Cybernetics, Part A: Systems and Humans

Quantified Score

Hi-index	0.00

Visualization

Abstract

This paper characterizes a genetic algorithm based on the analysis of the workload of its operators. Different granular parallel implementations of a genetic algorithm in the GPU architecture are compared against the correspondent sequential version. With the help of three benchmark problems, a complete characterization of the relative execution times of the genetic operators, varying the population cardinality and the genotype size, is offered. The best speedups, obtained with large populations, are higher than one thousand times faster than the corresponding sequential version. The assessment of different granularity levels shows that the two-dimensional parallelism supported by the GPU architecture is valuable for the crossover operator.