expl12_4.m

国外经典书籍MULTIVARIABLE FEEDBACK CONTROL-多变量反馈控制 的源码

% Example of a Generalized eigenvalue problem (upper bound on mu)
%
% Copyright 1996-2003 Sigurd Skogestad & Ian Postlethwaite
% $Id: Muup_LMI.m,v 1 2004/26/04 kariwala Exp $

% Delta: Structured matrix with 2 x 2 block and a repeated scalar 
M = randn(4,4);

setlmis([])

P = lmivar(1,[2 0;2 1]); % Structured matrix that commutes with Delta
gamma = lmivar(1,[1 1]);

Ppos = newlmi;                  % New LMI 
lmiterm([-Ppos 1 1 P],1,1)     % P > 0

MuupLMI = newlmi;                  % New LMI 

lmiterm([MuupLMI 1 1 P],M',M)  % M'PM (left hand side)
lmiterm([-MuupLMI 1 1 P],1,1)  % P (right hand side)

LMIsys = getlmis;               % Obtaining the system of LMIs

[gmin,xopt] = gevp(LMIsys,1);  % Solving the minimization problem
 
Popt = dec2mat(LMIsys,xopt,P);  % Obtaining the optimal P
Dopt = sqrtm(Popt);             % and scaling matrices