A mathematical model of task allocation of microseism data parallel computing is established in this paper. To overcome the shortcomings the traditional genetic algorithm has in low searching efficiency on complicated problems and easily falling into premature convergence, an integer code-based partheno-genetic algorithm is set up according to the practical application of task allocation in microseism data parallel computing, and the search capability of the algorithm is improved by optimising the design of mutation operator in the algorithm. Experiment results demonstrate that, the proposed algorithm has achieved considerable improvements with respect to the solution accuracy, global optimisation search and the prevention of premature convergence in comparison with traditional genetic algorithm.%建立了地脉动数据并行计算任务分配的数学模型,针对传统遗传算法对复杂问题搜索效率低、易陷入"早熟收敛"的缺点,根据地脉动数据并行计算中的任务分配问题的实际应用,构建了基于整数编码的单亲遗传算法,并通过优化设计算法中的变异算子提高了算法的搜索能力.实验结果表明,该算法在求解精度、全局寻优和防止过早收敛方面相对于传统遗传算法有了显著提高.
展开▼
