乘子交替方向法的一些收敛性质
发布时间:2018-08-14 19:51
【摘要】:正1引言本文讨论的两个可分离算子的线性约束凸优化问题是min{θ_i(x)+θ_2(y)|Ax+By=b,x∈X,y∈y},(1.1)其中A∈R~(m×n_1),B∈R~(m×n_2),b∈R~m;X?R~(n_1),y?R~(n_2)是闭凸集;θ_1(x):R~(n_1)→R和θ_2(y):R~(n_2)→R是(不一定光滑的)凸函数.这类问题大量出现在图像处理,机器学习等稀疏优化领域[2].乘子交替方向法(Alternating Directions Method of Multipliers),简称ADMM,通常称之为交替方向法,最初由Glowinski等为偏微分方程数值求解在[7,8],中
[Abstract]:The linear constrained convex optimization problem of two separable operators discussed in this paper is min {胃 I (x) 胃 2 (y) Ax Bybn x 鈭,
本文编号:2183945
[Abstract]:The linear constrained convex optimization problem of two separable operators discussed in this paper is min {胃 I (x) 胃 2 (y) Ax Bybn x 鈭,
本文编号:2183945
本文链接:https://www.wllwen.com/kejilunwen/yysx/2183945.html