Abstract: According to the problems of weak adaptability of single model, a method for CNC system software reliability estimation by using weighted combinational models was proposed. Firstly, three weighted combinational models, namely weighted arithmetic, weighted geometric, and weighted harmonic combinations, were proposed according to the characteristics of CNC system software. Secondly, the improved genetic algorithm was adopted to solve optimal weights for model combinations. Lastly, CNC system software failure data was taken into the models to estimate the faults of system, and then the estimation of CNC system software reliability was completed. Experiments were performed based on four sets of software failure data, and numerical results show that the weighted combinational model has a greater adaptability than that of single models. Then, the proposed models were applied to a self-developed handheld CNC system, and results verifies the feasibility and validity of the proposed method.

Key words: CNC system, software reliability, weighted combination, genetic algorithm

摘要：

针对单一模型适应度不高的问题，提出一种采用加权组合模型对数控系统软件进行可靠性评估的方法。首先根据数控系统软件的特点，分别建立算术加权、几何加权及调和加权三种组合模型；然后采用改进的遗传算法实现模型最优加权值的求解；最后将数控系统软件的失效数据代入模型，预估出了系统软件的错误情况，完成了对数控系统软件可靠性的估计。实验以四组软件失效数据为对象，实验结果表明加权组合模型比单一模型具有更高的适应性；将模型应用于自主研发的手持数控系统，验证了该方法的可行性和有效性。

关键词: 数控系统, 软件可靠性, 加权组合, 遗传算法