多步迭代正则化梯度法的收敛性

The Convergence of the Multi-step Iteratively Regularized Gradient Algorithm

分析一族多步迭代正则化梯度法的收敛性.分别利用修正的后验准则和先验准则,在修正源条件下得到算法关于误差的收敛性和收敛率结果.

The paper analyzed the convergence of a family of multi - step iteratively regularized algorithms for ill - posed operator equation. Employing a - priori rule and a - posteriori rule respectively, these algorithms are convergent and stable with respect to the operator error under the modified source condition.