串并联系统中支持实时替换的混合冗余策略优化

发布时间：2018-06-20 02:57

本文选题：冗余分配 + 可靠性优化　；参考：《软件学报》2017年02期

【摘要】：在需要长时间可靠运行的软件系统中,由于持续运行时间和任务响应速度的要求增加,工作组件在被探测到失效后将被冗余组件实时替换.但现有可靠性优化研究通常假设冷备份冗余在所有积极冗余组件失效后才使用.针对支持实时替换的混合冗余策略,对其冗余度优化分配进行研究.该策略不仅能够保障系统可靠性,而且能够保障系统性能,故选用实时可用性和任务完成效率两类约束条件,建立冗余配置代价最小化模型.基于马尔可夫链理论对可靠性及性能两类系统指标进行定量分析;采用数值计算方法对非线性的状态分析模型进行计算;改进二元组编码遗传算法对上述优化问题进行求解.采用实例对串并联系统中实时可用性及任务完成效率的分析进行了说明,并对优化冗余分配模型进行了验证.实验结果表明,在相同冗余度下,支持实时替换的混合冗余策略在任务完成效率方面优于传统的混合冗余策略.所以,在相同约束条件下不同混合冗余策略需要采用不同的冗余优化配置方案.
[Abstract]:In software systems that need to run reliably for a long time, due to the increasing requirements of continuous running time and task response speed, the work components will be replaced by redundant components in real time after detection of failure. However, existing reliability optimization studies generally assume that cold backup redundancy is only used after all active redundant components fail. Aiming at the hybrid redundancy strategy which supports real-time substitution, the optimal allocation of redundancy degree is studied. The strategy can not only guarantee system reliability but also guarantee system performance. Therefore, two kinds of constraints, real-time availability and task completion efficiency, are selected to minimize the cost of redundant configuration. Based on Markov chain theory, the reliability and performance of the two kinds of system indicators are quantitatively analyzed; the nonlinear state analysis model is calculated by numerical method; and the optimization problem is solved by improved binary coded genetic algorithm. An example is given to illustrate the real-time availability and task completion efficiency of series-parallel systems, and the optimal redundancy allocation model is verified. The experimental results show that the hybrid redundancy strategy which supports real-time replacement is superior to the traditional hybrid redundancy strategy in task completion efficiency under the same redundancy degree. Therefore, different mixed redundancy strategies need to adopt different redundant optimal configuration schemes under the same constraint conditions.
【作者单位】：中国科学院重庆绿色智能技术研究院;
【基金】：国家自然科学基金(61309005) 重庆市前沿与应用基础研究计划(cstc2014jcyj A40015)~~
【分类号】：TP18;TB114.3

【相似文献】