-
Notifications
You must be signed in to change notification settings - Fork 61
/
proj_max.m
36 lines (32 loc) · 1.06 KB
/
proj_max.m
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
function op = proj_max( q )
%PROJ_max Projection onto the scaled max-function ball.
% OP = PROJ_MAX( Q ) returns an operator implementing the
% indicator function for the max-function ball of size q,
% { X | max( X(:) ) <= q }. Q is optional; if omitted,
% Q=1 is assumed. But if Q is supplied, it must be a
% real scalar (negative is OK).
% Dual: prox_l1pos.m
% See also: prox_l1pos, prox_linf, proj_l1, proj_linf
if nargin == 0,
q = 1;
elseif ~isnumeric( q ) || ~isreal( q ) || numel( q ) ~= 1
error( 'Argument "q" must be a real scalar.' );
end
op = @(varargin)proj_linf_q( q, varargin{:} );
function [ v, x ] = proj_linf_q( q, x, t )
v = 0;
switch nargin,
case 2,
if nargout == 2,
error( 'This function is not differentiable.' );
elseif max( x(:) ) > q,
v = Inf;
end
case 3,
x = min( x, q );
otherwise,
error( 'Not enough arguments.' );
end
% TFOCS v1.3 by Stephen Becker, Emmanuel Candes, and Michael Grant.
% Copyright 2013 California Institute of Technology and CVX Research.
% See the file LICENSE for full license information.