FBF_PRIMAL_DUAL - forward backward forward primal dualUsagesol = fbf_primal_dual(x_0,f1, f2, f3, param); sol = fbf_primal_dual(x_0,f1, f2, f3); [sol,info,objective] = fbf_primal_dual(...); Input parameters
Output parameters
Descriptionfbf_primal_dual (using forward backward forward based primal dual) solves:
\begin{equation*}
sol = \min_x f_1(x) + f_2( L x) + f_3(x)
\end{equation*}
where \(x\) is the optimization variable with \(f_1\) or \(f_3\) a smooth function and \(L\) a linear operator. \(f_1\) and \(f_3\) are defined like other traditional functions. Note that f2 is a structure of a functions with:
Optionally you can define
param a Matlab structure containing solver paremeters. See the function solvep for more information. Additionally it contains those aditional fields:
References:N. Komodakis and J.-C. Pesquet. Playing with duality: An overview of recent primal-dual approaches for solving large-scale optimization problems. arXiv preprint arXiv:1406.5429, 2014. |