考虑随机权重阈值的多态k/N系统任务可用度建模与优化

在权重阈值随机变化的条件下,现有多态系统可用度建模在实际工程应用中具有局限性.本文引入数量阈值并考虑权重阈值随机性,分析多态k/N系统中部件数量和权重不能彼此映射的问题,提出新的随机联合运算符结合通用生成函数的方法,建立系统任务可用度评价模型.本方法适用于权重阈值服从连续型随机分布的情形,优于传统递归算法.然后,分别构建系统可用性约束或成本约束下的系统优化模型,通过遗传算法编程,求解部件各状态的最优概率及权重.最后,以船舶运输系统为例阐述数量阈值对系统可用度的影响,并给出随机权重阈值下多态系统的设计优化结果.研究成果可为系统工程师开展可靠性设计和改进提供方法依据.

Mission-availability modeling and optimization of multi-state k/N systems under random weight threshold

Under the condition of randomness of weight threshold, the existing availability modeling of multi-state systems has limitations in practice. In this paper, quantity threshold and randomness of the weight threshold are both considered, and a new method of combining stochastic joint operator with general generation function is proposed to solve the problem that the quantity and weight of the components in the k/N systems cannot be mapped to each other. Thus, the mission-availability evaluation model of the system is obtained, which is applicable to the case where the weight threshold is subject to continuous random distribution, and which is superior to the traditional recursive algorithm. Then, two system optimization models under system availability constraint or system cost constraint are constructed respectively, and the genetic algorithm programming is applied to achieve the optimal component's probabilities and weights. Finally, the ship transportation is taken as an example to illustrate the effect of the quantity threshold on system mission-availability and the optimization results under the random weight threshold are given. The research results can be useful for solutions and methods for system engineers to carry out reliability design and improvement.