Download the zip file of cvx
http://cvxr.com/cvx/download/ by downloading cvx-w64.zip
Require a licence for academic use
a file named cvx_license.dat
Install with licence
unzip the zip file into the toolbox directory ($MatlabRoot oolbox)
run the command in matlab after entering the directory of cvx (adding the path of licence file)
>> cvx_setup E:softwareMATLABR2014a oolboxcvxcvx_license.dat
Then
---------------------------------------------------------------------------
CVX: Software for Disciplined Convex Programming (c)2014 CVX Research
Version 2.1, Build 1112 (ef41608) Sun Oct 2 17:15:09 2016
---------------------------------------------------------------------------
Test with simple script
cvx_test.m
1 m = 20; n = 10; p = 4; 2 A = randn(m,n); b = randn(m,1); 3 C = randn(p,n); d = randn(p,1); e = rand; 4 cvx_begin 5 variable x(n) 6 minimize( norm( A * x - b, 2 ) ) 7 subject to 8 C * x == d 9 norm( x, Inf ) <= e 10 cvx_end
run the script
1 >> cvx_test 2 3 Calling SDPT3 4.0: 56 variables, 22 equality constraints 4 For improved efficiency, SDPT3 is solving the dual problem. 5 ------------------------------------------------------------ 6 7 num. of constraints = 22 8 dim. of socp var = 41, num. of socp blk = 11 9 dim. of linear var = 11 10 dim. of free var = 4 *** convert ublk to lblk 11 ******************************************************************* 12 SDPT3: Infeasible path-following algorithms 13 *******************************************************************
------------------------For abbreviatin --------------
14 rel. primal infeas (scaled problem) = 2.01e-09 15 rel. dual " " " = 6.95e-10 16 rel. primal infeas (unscaled problem) = 0.00e+00 17 rel. dual " " " = 0.00e+00 18 norm(X), norm(y), norm(Z) = 1.5e+00, 4.5e+00, 6.4e+00 19 norm(A), norm(b), norm(C) = 1.8e+01, 2.0e+00, 6.3e+00 20 Total CPU time (secs) = 0.07 21 CPU time per iteration = 0.01 22 termination code = 0 23 DIMACS: 2.0e-09 0.0e+00 1.2e-09 0.0e+00 6.3e-09 6.2e-09 24 ------------------------------------------------------------------- 25 26 ------------------------------------------------------------ 27 Status: Solved 28 Optimal value (cvx_optval): +4.29018