| 摘要: |
| 文章讨论了并行遗传算法的迁移现象及群体规模估算模型 ,分析了迁移的过程,揭示了迁移的实质,并提出了在理想条件下的迁移计算模型.基于迁移计 算模型导出了粗粒度并行遗传算法进化质量估量模型.实验结果表明,模型比较准确地预测了 在特定领域进化质量的变化趋势. |
| 关键词: 粗粒度并行遗传算法,积木块,迁移模式,子群体连接拓扑 结构,群体规模估算模型. |
| DOI： |
| 分类号: |
| 基金项目: |
|
| Modeling Idealized Parallel Genetic Algorithms |
|
HOU Guang-kun,LUO Jiang-peng
|
| Abstract: |
| The migration and population sizing of paralle l genetic algorithms are discussed in this paper, the migration procedure is ana lysed, the nature of migration is revealed, then a computation model of migratio n in the ideal conditions is propose. Base on this model, a quality estimation f or parallel genetic algorithms with large grain is drived, it predicts the quali ty tendency of evolution quite precisely in special domain, by experiments. |
| Key words: Parallel genetic algorithms with large grain, building blocks convergence, migra tion schemes, communication topology, population sizing. |
