模糊可靠性计算问题最后也要转化为常规可靠性优化来进行计算,但目前所用的手段多基于传统的最优化理论中优化算法:如复合型法、惩罚函数法等,这些方法存在局部极值和对目标函数的可微性有严格要求的苛刻条件,并且优化结果与初始值有较大的相关性等难以克服的缺点。在Matlab环境下,应用GA工具箱,对圆柱螺旋弹簧进行优化设计,旨在克服常规算法的缺点,得到全局最优解。研究应用结果表明在Maltab环境下,遗传算法工具箱使用简单方便,可直接应用库函数文件,且能根据不同的优化指标、计算精度寻找出最佳结果,充分体现了遗传算法的全局搜索性能。该优化方法具有计算可靠,快捷高效和图形结果可视化等特点,并具有广阔的应用前景。
The problem of fuzzy reliability should be calculated in the form of conventional reliability optimization. At present, the algorithms of complex and penalty function are applied as the computing method. These algorithms have some rigorous terms, such as the function must be differentiable, and the result has much relation with the initial value, The research is carried out using the GA toolbox in Matlab. The genetic algorithm is applied for overcoming the shortcomings of conventional algorithms. It makes the mechanical fuzzy reliability optimal design of cylindrical spring as a study example. The study shows that GA Toolbox in Matlab is convenient to be used and the optimal results can be got according to different optimization indexes and different calculation precisions, thus the global search function of GA is exhibited sufficiently. The optimization method has the characteristics such as reliable calculation, high efficiency and visualized graphic results etc., and has a broad application prospect as well.