how to accelate the 'for loop' command' It's not end processing !!

1 view (last 30 days)
clear all, close all, clc
%import data from excel sheet I have attached
x = x.'; %just x variable
n =5000; % number of rows
m =2726; % number of columns
index1 = 1:n;
index2 = n:n+m-1;
X = []; Xprime=[];
for ir = 1:size(x,1)
% Hankel blocks ()
c = x(ir,index1).'; r = x(ir,index2);
H = hankel(c,r).';
c = x(ir,index1+1).'; r = x(ir,index2+1);
UH= hankel(c,r).';
X=[X,H]; Xprime=[Xprime,UH];
end
%% DMD of x
dt = 1;
t=[0:1:7725];
r = 16 ;
[U, S, V] = svd(X, 'econ');
r = min(r, size(U,2));
U_r = U(:, 1:r); % truncate to rank-r
S_r = S(1:r, 1:r);
V_r = V(:, 1:r);
Atilde = U_r' * Xprime * V_r / S_r; % low-rank dynamics
[W_r, D] = eig(Atilde);
Phi = Xprime * V_r / S_r * W_r; % DMD modes
%PhiP = U_r * V_r; % Projected DMD modes
lambda = diag(D); % discrete-time eigenvalues
omega = log(lambda)/dt; % continuous-time eigenvalues
% Compute DMD mode amplitudes b
x1 = X(:, 2);
b = Phi\x1; % equal to b = pinv(Phi)*x1
% DMD reconstruction
time_dynamics = zeros(r, length(t));
for iter = 1:length(t)
time_dynamics(:,iter) = (b.*exp(omega*t(iter)));
end
Xdmd = Phi * time_dynamics;
% DMD reconstruction
rr = length(lambda) ;
T = size(X,2) ;
time_dmd = zeros(T-1,rr);
for iter = 1:T-1
for p = 1:rr
time_dmd(iter,p) = b(p)*(exp(omega(p)*t(iter)));
Xdm(:,iter,p) = real(Phi(:,p)*(b(p).*exp(omega(p)*t(iter))));
end
end
X_dmd = sum(Xdm,3) ;

Accepted Answer

DGM
DGM on 30 Jun 2023
Edited: DGM on 30 Jun 2023
Again, you're trying to construct large (~1.7 GB) arrays by growing them (in the reconstruction part). That will take an impractically long time (at least an hour or so). Like the other answer I gave, the loops are unnecessary anyway, so the whole issue can be avoided. Even on my dumpster computer with minimal ram, the loopless replacement section takes about 5 seconds.
As to whether the rest of the code works, I don't know. That's all over my head. I just tried to make sure my edits didn't change the results.
n = 5000; % number of rows
m = 2726; % number of columns
x = rand(1,n+m);
index1 = 1:n;
index2 = n:n+m-1;
% preallocate
szx = size(x,1);
X = zeros(m,n*szx);
Xprime = zeros(m,n*szx);
tic
for ir = 1:szx
% Hankel blocks ()
c = x(ir,index1).';
r = x(ir,index2);
H = hankel(c,r).';
c = x(ir,index1+1).';
r = x(ir,index2+1);
UH = hankel(c,r).';
cols = n*(ir-1)+1:n*ir;
X(:,cols) = H;
Xprime(:,cols) = UH;
end
toc
% preallocating would have been more important if x had more than one row
% if x never has more than one row, then the entire loop is unnecessary
%% DMD of x
dt = 1;
t = 0:1:7725;
r = 16 ;
[U, S, V] = svd(X, 'econ');
%%
r = min(r, size(U,2));
U_r = U(:, 1:r); % truncate to rank-r
S_r = S(1:r, 1:r);
V_r = V(:, 1:r);
Atilde = U_r' * Xprime * V_r / S_r; % low-rank dynamics
[W_r, D] = eig(Atilde);
Phi = Xprime * V_r / S_r * W_r; % DMD modes
%PhiP = U_r * V_r; % Projected DMD modes
lambda = diag(D); % discrete-time eigenvalues
omega = log(lambda)/dt; % continuous-time eigenvalues
% Compute DMD mode amplitudes b
x1 = X(:, 2);
b = Phi\x1; % equal to b = pinv(Phi)*x1
% DMD reconstruction
time_dmd = b.'.*exp(omega.'.*reshape(t(1:size(X,2)-1),[],1));
Xdm = permute(Phi,[3 2 1]).*time_dmd;
Xdm = real(permute(Xdm,[3 1 2]));
X_dmd = sum(Xdm,3);

More Answers (0)

Categories

Find more on Programming in Help Center and File Exchange

Community Treasure Hunt

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

Start Hunting!