CVX: quickstart2 Bound-constrained Least squares
Least squaresに制約条件を追加する．今回は上界と下界を
minimize ||Ax-b||2
subject to l<=x<=u
- m = 16; n = 8;
- A = randn(m,n); b = randn(m,1);
- bnds = randn(n,2);
- l = min( bnds, , 2 ); u = max( bnds, , 2 );
- cvx_begin
- variable x(n);
- minimize( norm(A*x-b) );
- subject to
- x >= l;
- x <= u;
- cvx_end
本当にl<=x<=uとなっているか確認．
>> [l x u]
ans =
-2.1237 -0.1904 0.4494
-0.5046 -0.5046 0.1006
-1.2706 0.4633 0.8261
-0.3826 -0.3826 0.5362
0.6487 0.6487 0.8979
-0.1319 -0.1319 0.8257
-1.0149 -0.1472 -0.1472
-0.4711 0.1281 1.0078
制約なしLeast squaresで求めた解は
>> x_ls
x_ls =
0.1768
-0.6757
-0.7052
-0.7581
-0.2865
-0.6379
0.0581
-0.4439
となっており，確かに制約通りの解が求まっている事が分かる．
