摘要： 结合诊断模型与遗传算法，提出一种针对t-可诊断多处理器系统的遗传诊断算法。通过运用模型的诊断性质提高初始种群的质量，并定义判断结点状态的诊断方程，设计基于该方程的适应度函数。理论分析和模拟实验表明，算法具有较好的最优解搜索能力和较快的收敛速度。

关键词: t-可诊断系统, 容错, 诊断模型, 遗传算法, 诊断算法

Abstract: Aiming at t-diagnosable multiprocessor system, this paper proposes a genetic diagnosis algorithm combining genetic algorithm and the diagnosis model. By applying the diagnosis property of the model, the quality of initial population is improved. A set of equations that estimate the statuses of the nodes in a system is proposed. Based on this, a novel fitness function is presented. Both theoretical analysis and simulation results indicate that it is a convergent and efficient algorithm.

Key words: t-diagnosable system, fault tolerance, diagnosis model, genetic algorithm, diagnosis algorithm

中图分类号: 

• TP306