并行遗传算法将并行计算机的高速并行性和遗传算法固有的并行性相结合，极大地提升了遗传算法的求解速度和质量。在主从式0细粒度和粗粒度这三类遗传算法并行化模型中，粗粒度模型以其较小的通讯开销和对种群多样化，获得了最广泛的应用。本文概括了基于模式定理和有限状态马尔可夫链的遗传算法理论，总结了前人在粗粒度模型下开展的理论分析和实践应用，并指出并行遗传算法的研究将向异步化，理论化和模型化的方向发展-而有限状态马尔可夫链是构建并行遗传算法可执行模型的有力工具。-Parallel genetic algorithm parallel computers and high-speed parallel genetic algorithm combining the inherent parallelism, which greatly enhanced the genetic algorithm to solve the speed and quality. In the master-slave 0 grained and coarse-grained parallel genetic algorithm to these three models, the coarse-grained model with its smaller communication overhead and population diversity, access to the most widely used. This paper summarizes the pattern-based theorem and finite state Markov chain genetic algorithm theory, summed up in the coarse-grained model predecessors carried out under the theoretical analysis and practical application, and that will be parallel genetic algorithm based asynchronous, theoretical and model of direction- and finite state Markov chain is to build a parallel genetic algorithm executable model a powerful tool.