RMS算法的收敛性

convergence of RMS Algorithm

摘要: RMS算法是K.S.Chandra等人1982年提出的一种可用于非光滑函数的线搜索优化算法¹它具有调用函数次数少,CPU时间短等突出优点。
本文补充了文献¹中未讨论的退化情况,并对RMS算法加以修改,使其理论上趋于完善。本文还论证了在一定条件下RMS算法的收敛性。

Abstract: The RMS algorithm was initiated by K.S.Chandra and M.V.Rao in1. But in1 the possiblity of degeneration was neglected.In our paper,we construct a counterexample showing that degeneration may infer the non-convergence of the algorithm and we make some corrections to the algorithm so that we can prove the convergence of the corrected algorithm.