基于改进可行方向法和随机方向法的V带传动优化

结合可行方向法和随机方向法2种优化方法的优点,提出改进的优化方法用于V带设计,首先利用随机方向法产生初始点,当随机方向法产生随机方向时则以可行方向法产生可行方向的条件进行约束,对随机单位矢量进行甄选,只有同时符合可行条件和下降条件的单位矢量方可作为可行搜索方向,在沿选定搜索随机方向进行搜索时利用梯度法求导确定最佳步长,避免加速步长法的重复迭代。改进后的优化方法结合两者的优点并克服相应的不足,减少产生随机单位矢量的数量,提高了优化效率。计算实例表明改进的方法可靠、有效。

Optimization of V-belt Drive Based on the Improved Feasible Direction Method and Random Direction Method

Combining the advantages of feasible direction optimization method and random direction optimization method, an improved method for the v-belt design was raised. First, using random direction method to generate a initial point, when the random direction method generated a random direction, feasible direction method placed the conditions arising feasible direction constraints for the selection of the random unit vector. Only the unit vector complied with the viable and the decline condition, it could be a practical search direction. The derivation of the gradient method was used to determine the optimum step size when searching along the selected direction and to avoid repeated iterations of accelerated step method. Improved optimization method combined the advantages of both and overcame the ｜ack of appropriate, and reduced the number of random unit vector and improved optimize efficiency. The example calculation shows that the improved method is reliable and effective.