Solving linear systems in Matlab

1 view (last 30 days)
L'O.G.
L'O.G. on 24 Aug 2023
Edited: Bruno Luong on 24 Aug 2023
How do I solve for x in u = A(x+b)? A is a matrix, and the other terms are vectors. Isn't it something like x = A\u - b ? I'm not sure about subtracting the b term, however.
  1 Comment
Bruno Luong
Bruno Luong on 24 Aug 2023
Edited: Bruno Luong on 24 Aug 2023
You need to be careful when the system is undedetermined (usually more unknown than equation)
In some sense x = A\u - b is the worse you can pick.

Sign in to comment.

Accepted Answer

Torsten
Torsten on 24 Aug 2023
Edited: Torsten on 24 Aug 2023
x = A\(u-A*b)
But it's the same as
x = A\u - b
  1 Comment
Bruno Luong
Bruno Luong on 24 Aug 2023
Edited: Bruno Luong on 24 Aug 2023
Is it?
A=rand(3,4);
b=rand(4,1);
u=rand(3,1);
% 4 different solutions
x1=A\(u-A*b)
x1 = 4×1
-0.2210 0 -0.6049 -0.1026
x2=A\u-b
x2 = 4×1
-0.0231 -0.8875 -1.3122 1.0476
x3=lsqminnorm(A,u)-b
x3 = 4×1
-0.1409 -0.3592 -0.8912 0.3630
x4=lsqminnorm(A,u-A*b)
x4 = 4×1
-0.2409 0.0891 -0.5339 -0.2181
% All give the same "fitness"
A*x1-u
ans = 3×1
-0.9852 -1.3140 -1.3694
A*x2-u
ans = 3×1
-0.9852 -1.3140 -1.3694
A*x3-u
ans = 3×1
-0.9852 -1.3140 -1.3694
A*x4-u
ans = 3×1
-0.9852 -1.3140 -1.3694
norm(x1)
ans = 0.6522
norm(x2)
ans = 1.8993
norm(x3)
ans = 1.0368
norm(x4) % always the smallest
ans = 0.6314

Sign in to comment.

More Answers (1)

Steven Lord
Steven Lord on 24 Aug 2023
Define a new variable y such that y = x+b.
Using this new variable simplifies your system to u = A*y.
Solve this simplified system for y using the backslash operator, y = A\u.
Substituing into that solution using the definition of y, we know x+b = A\u.
Subtract b from both sides to get x = (A\u)-b.

Categories

Find more on Mathematics and Optimization in Help Center and File Exchange

Products


Release

R2021b

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!