考虑求解非光滑方程组的三次正则化方法及其收敛性分析.利用信赖域方法的技巧，保证该方法是全局收敛的.在子问题非精确求解和BD正则性条件成立的前提下，分析了非光滑三次正则化方法的局部收敛速度.最后，数值实验结果验证了该算法的有效性.
A cubic regularization method and its convergence for solving a nonsmooth system of equations are studied in this paper. By applying the classical trust region technique, the proposed method is ensured to be globally convergent. When BDregular condition is satisfied and the subproblem is inexactly solved, we analyze the local convergence rate of the nonsmooth cubic regularization method. Finally, the efficiency of our method is verified by numerical results.
[1] Pang J S, Qi L Q. Nonsmooth equations:motivation and algorithms[J]. SIAM Journal on Optimization, 1993, 3(3):443-465.
[2] Qi L Q, Sun D F. A survey of some nonsmooth equations and smoothing Newton methods[M]//Progress in Optimization, New York:Springer, 1999.
[3] Nesterov Y, Polyak B T. Cubic regularization of Newton's method and its global performance[J]. Mathematical Programming, 2006, 108(1):177-205.
[4] Cartis C, Gould N I M, Toint P L. On the complexity of steepest descent, Newton's and regularized Newton's methods for nonconvex unconstrained optimization[J]. SIAM Journal on Optimization, 2010, 20(6):2833-2852.
[5] Cartis C, Gould N I M, Toint P L. Adaptive cubic regularisation methods for unconstrained optimization. Part I:motivation, convergence and numerical results[J]. Mathematical Programming, 2011, 127(2):245-295.
[6] Cartis C, Gould N I M, Toint P L. Adaptive cubic overestimation methods for unconstrained optimization. Part Ⅱ:worst-case function-evaluation complexity[J]. Mathematical Programming, 2011, 130(2):295-319.
[7] Qi L Q, Sun J. A nonsmooth version of Newton's method[J]. Mathematical Programming, 1993, 58(1):353-367.
[8] Facchinei F, Pang J S. Finite-Dimensional Variational Inequalities and Complementarity Problems I[M]. New York:Springer-Verlag, 2003.
[9] Facchinei F, Pang J S. Finite-Dimensional Variational Inequalities and Complementarity Problems Ⅱ[M]. New York:Springer-Verlag, 2003.
[10] Lukšan L. Inexact trust region method for large sparse systems of nonlinear equations[J]. Journal of Optimization Theory and Applications, 1994, 81(3):569-590.
[11] Qi L Y, Xiao X T, Zhang L W. A parameter-self-adjusting Levenberg-Marquardt method for solving nonsmooth equations[J]. Journal of Computational Mathematics, 2016, 34(3):317-338.
