1 | <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN"> |
---|
2 | <html> |
---|
3 | |
---|
4 | <head> |
---|
5 | <meta http-equiv="Content-Language" content="en-us"> |
---|
6 | <title>YALMIP Example : Complex-valued problems</title> |
---|
7 | <meta http-equiv="Content-Type" content="text/html; charset=windows-1251"> |
---|
8 | <meta content="Microsoft FrontPage 6.0" name="GENERATOR"> |
---|
9 | <meta name="ProgId" content="FrontPage.Editor.Document"> |
---|
10 | <link href="yalmip.css" type="text/css" rel="stylesheet"> |
---|
11 | <base target="_self"> |
---|
12 | </head> |
---|
13 | |
---|
14 | <body leftMargin="0" topMargin="0"> |
---|
15 | |
---|
16 | <div align="left"> |
---|
17 | |
---|
18 | <table border="0" cellpadding="4" cellspacing="3" style="border-collapse: collapse" bordercolor="#000000" width="100%" align="left" height="100%"> |
---|
19 | <tr> |
---|
20 | <td width="100%" align="left" height="100%" valign="top"> |
---|
21 | |
---|
22 | <h2>Complex-valued problems</h2> |
---|
23 | <hr noShade SIZE="1"> |
---|
24 | <p>YALMIP supports complex valued constraints for all solvers by |
---|
25 | automatically converting complex-valued problems to real-valued problems.</p> |
---|
26 | <p>To begin with, let us just define a simple linear complex problem to |
---|
27 | illustrate how complex variables and constraints are generated and |
---|
28 | interpreted.</p> |
---|
29 | <table cellPadding="10" width="100%"> |
---|
30 | <tr> |
---|
31 | <td class="xmpcode"> |
---|
32 | <pre>p = sdpvar(1,1,'full','complex'); % A complex scalar (4 arguments necessary) |
---|
33 | s = sdpvar(1,1)+sqrt(-1)*sdpvar(1,1); % Alternative definition |
---|
34 | F = set('0.9>imag(p)'); % Imaginary part constrained |
---|
35 | F = F+set('0.01>real(p)'); % Real part constrained |
---|
36 | |
---|
37 | F = F+set('0.1+0.5*sqrt(-1)>p'); % Both parts constrained |
---|
38 | |
---|
39 | F = F+set('s+p==2+4*sqrt(-1)'); % Both parts constrained</pre> |
---|
40 | </td> |
---|
41 | </tr> |
---|
42 | </table> |
---|
43 | <p>To see how complex-valued constraints can be used in a more advanced |
---|
44 | setting, we solve the covariance estimation problem from the <a href="solvers.htm#sedumi">SeDuMi</a> |
---|
45 | manual. The problem is to find a positive-definite Hermitian Toeplitz matrix |
---|
46 | <b>Z</b> such that the Frobenious norm of <b><font face="Tahoma">P-Z</font></b> is minimized (<b><font face="Tahoma">P</font></b> is a given complex |
---|
47 | matrix.)</p> |
---|
48 | <p>The matrix <b>P</b> is</p> |
---|
49 | <table cellPadding="10" width="100%"> |
---|
50 | <tr> |
---|
51 | <td class="xmpcode"> |
---|
52 | <pre>P = [4 1+2*i 3-i;1-2*i 3.5 0.8+2.3*i;3+i 0.8-2.3*i 4];</pre> |
---|
53 | </td> |
---|
54 | </tr> |
---|
55 | </table> |
---|
56 | <p>We define a complex-valued Toeplitz matrix of the corresponding dimension</p> |
---|
57 | <table cellPadding="10" width="100%"> |
---|
58 | <tr> |
---|
59 | <td class="xmpcode"> |
---|
60 | <pre>Z = sdpvar(3,3,'toeplitz','complex')</pre> |
---|
61 | </td> |
---|
62 | </tr> |
---|
63 | </table> |
---|
64 | <p>A complex Toeplitz matrix is not Hermitian, but we can make it Hermitian |
---|
65 | if we remove the imaginary part on the diagonal.</p> |
---|
66 | <table cellPadding="10" width="100%"> |
---|
67 | <tr> |
---|
68 | <td class="xmpcode"> |
---|
69 | <pre>Z = Z-diag(imag(diag(Z)))*sqrt(-1);</pre> |
---|
70 | </td> |
---|
71 | </tr> |
---|
72 | </table> |
---|
73 | <p>Minimizing the Frobenious norm of <b><font face="Tahoma">P-Z</font></b> can be cast as minimizing the |
---|
74 | Euclidean norm of the vectorized difference <b><font face="Tahoma">P(:)-Z(:)</font></b>. By using a Schur |
---|
75 | complement, we see that this can be written as the following SDP.</p> |
---|
76 | <table cellPadding="10" width="100%"> |
---|
77 | <tr> |
---|
78 | <td class="xmpcode"> |
---|
79 | <pre>e = P(:)-Z(:) |
---|
80 | t = sdpvar(1,1); |
---|
81 | F = set(Z>0); |
---|
82 | F = F+set([t e';e eye(9)]>0); |
---|
83 | solvesdp(F,t);</pre> |
---|
84 | </td> |
---|
85 | </tr> |
---|
86 | </table> |
---|
87 | <p>The problem can be implemented more efficiently using a second order cone |
---|
88 | constraint.</p> |
---|
89 | <table cellPadding="10" width="100%"> |
---|
90 | <tr> |
---|
91 | <td class="xmpcode"> |
---|
92 | <pre>e = Z(:)-P(:) |
---|
93 | t = sdpvar(1,1); |
---|
94 | F = set(Z>0); |
---|
95 | F = F+set(cone(e,t)); |
---|
96 | solvesdp(F,t);</pre> |
---|
97 | </td> |
---|
98 | </tr> |
---|
99 | </table> |
---|
100 | <p>...or by using a quadratic objective function</p> |
---|
101 | <table cellPadding="10" width="100%"> |
---|
102 | <tr> |
---|
103 | <td class="xmpcode"> |
---|
104 | <pre>e = Z(:)-P(:) |
---|
105 | F = set(Z>0); |
---|
106 | solvesdp(F,e'*e);</pre> |
---|
107 | </td> |
---|
108 | </tr> |
---|
109 | </table> |
---|
110 | </td> |
---|
111 | </tr> |
---|
112 | </table> |
---|
113 | |
---|
114 | </div> |
---|
115 | |
---|
116 | </body> |
---|
117 | |
---|
118 | </html> |
---|