A Parallel Genetic Algorithm Based on Adaptive Migration Strategy

摘要

Genetic Algorithm is a modern artificial intelligence algorithm, which results from the process of simulation biological evolutionary. The parallelism of Genetic Algorithm has high performance in solving complex, large-scale, non-linear, non-differentiable optimization problems. This paper analyzes limitations of traditional parallel genetic algorithms, for its migration fixed blindness and other disadvantages. Adaptive migration strategy of parallel genetic algorithms (AMPGA) was proposed here, which was suitable for running on multi-core computers. This method combines the Genetic Algorithm with current computer architecture, which makes the Parallel Genetic Algorithm improves the convergence speed, fully exploit computing capability of computers. Experiments showed that AMPGA not only has faster convergent speed but also has higher precision and parallel efficiency.