function varargout = sumabsk(varargin) % SUMABSK Returns sum of k largest (by magnitude) (eigen-)values. % % s = SUMABSK(X,k) % % For a real vector X, SUMABSK returns the sum of the k largest (by magnitude) elements. % % For an Hermitian matrix X, SUMABSK returns the sum of the k largest (by magnitude) eigen-values. % % See also SUMK % *************************************************** % This file defines a nonlinear operator for YALMIP % % It can take three different inputs % For double inputs, it returns standard double values % For sdpvar inputs, it genreates a an internal variable % When first input is 'model' it generates the epigraph % % *************************************************** switch class(varargin{1}) case 'double' % What is the numerical value of this argument (needed for displays etc) if nargin == 1 error('sumabsk needs two arguments'); else X = varargin{1}; [n,m] = size(X); if (min(n,m) > 1 & ~ishermitian(X)) | (n~=m & ~isreal(X)) error('sumabsk can only be applied on real vectors and Hermitian matrices'); else k = min(length(X),varargin{2}); if min(n,m)==1 sorted = sort(abs(X)); else sorted = sort(abs(eig(X))); end varargout{1} = sum(sorted(max(1,end-k+1):end)); end end case 'sdpvar' % Overloaded operator for SDPVAR objects. Pass on args and save them. X = varargin{1}; [n,m] = size(X); if (min(n,m) > 1 & ~ishermitian(X)) | (n~=m & ~isreal(X)) error('sumabsk can only be applied on real vectors and Hermitian matrices'); else if nargin < 2 error('sumabsk needs two arguments'); else varargout{1} = yalmip('define',mfilename,varargin{:}); end end case 'char' % YALMIP sends 'model' when it wants the epigraph or hypograph if isequal(varargin{1},'graph') t = varargin{2}; % Second arg is the extended operator variable X = varargin{3}; % Third arg and above are the args user used when defining t. k = min(varargin{4},length(X)); [n,m] = size(X); Z = sdpvar(n,m); s = sdpvar(1,1); if min(n,m)==1 varargout{1} = (t-k*s-sum(Z) >= 0) + (Z >= 0) + (Z+s >= X >= -Z-s); else varargout{1} = (t-k*s-trace(Z) >= 0) + (Z >= 0) + (Z+s*eye(n) >= X >= -Z-s*eye(n)); end varargout{2} = struct('convexity','convex','monotonicity','none','definiteness','none','model','graph'); varargout{3} = X; else end otherwise end