Too many output arguments (five on output)

1 vue (au cours des 30 derniers jours)
Dmitry Surov
Dmitry Surov le 11 Juil 2022
Commenté : Les Beckham le 12 Juil 2022
Code below. Getting problems with "too many output", that's strange, cause cub_def has the same (5) output variables. Tried to do it with Fuction [a_spline, a_coeff_spline]=cub_def(x, a, x, 0, 0); and then call just
[a_spline, a_coeff_spline]=cub_def(x, a, x, 0, 0) - didn't help
clc;
close all;
imtool close all;
clear;
workspace;
% Read data
[filename, filepath]=uigetfile;
data_from_file=dlmread([filepath, filename],';', 3, 0);
% Read data
x=data_from_file(:,1);
y=data_from_file(:,2);
%Max Iterations
i_max=10000;
%Reg coef limits
A_min=-3;
A_max=3;
alpha_min=0.25;
alpha_max=1;
f_min=0.158;
f_max=1;
%Temperature
t_min=0;
t_max=1;
%Temperature decrease law
denom=1:1:i_max-1;
t=t_max./denom;
t(length(t)+1)=t_min;
% initial approximation
b_pribl_anneal=[295, 140, 7.559];
%array for calculating the residual function at the i-th iteration
sigma=zeros(i_max, 1);
%array with residual function for response
sigma_goal=0.5*sum((b_pribl_anneal(1).*sin(2*pi*b_pribl_anneal(3).*x)./(b_pribl_anneal(2)^2+x.^2))-y.^2).*ones(size(sigma));
%calculate the first value of the residual function by substituting
%the coefficients from the initial approximation
sigma_goal=0.5*sum((b_pribl_anneal(1).*sin(2*pi*b_pribl_anneal(3).*x)./(b_pribl_anneal(2)^2+x.^2))-y.^2);
%create fields for charts and give them names
plot_anneal_figure=figure();
for i=1:i_max;
if mod(i, 10000)==0 || i==1;
%build a graph with experimental points and the calculated model
figure(plot_anneal_figure);
plot(x, y, 'LineStyle', 'none', 'Marker', '*');
hold on;
%solution at the current iteration
plot(x, b_pribl_anneal(1).*sin(2*pi*b_pribl_anneal(3).*x)./(b_pribl_anneal(2)^2+x.^2), 'LineWidth', 2, 'Color', 'r') ;
legend('Experimental data', ['Simulated annealing', num2str(i) 'iterations']);
hold off;
end
%calculate sigma at the current iteration
sigma(i)=0.5*sum((b_pribl_anneal(1).*sin(2*pi*b_pribl_anneal(3).*x)./(b_pribl_anneal(2).^2+x.^2))-y.^2);
%generate candidate for solution
b_shift=[rand()*(A_max-A_min) rand()*(alpha_max-alpha_min) rand()*(f_max-f_min)]*(-1)^randi([0 1])*t(i);
b_pribl_anneal_new=b_pribl_anneal+b_shift;
%check if we stay within the allowed range
if b_pribl_anneal_new(1)>A_max || b_pribl_anneal_new(1)<A_min;
b_pribl_anneal_new(1)=b_pribl_anneal(1);
end
if b_pribl_anneal_new(2)>alpha_max || b_pribl_anneal_new(2)<alpha_min;
b_pribl_anneal_new(2)=b_pribl_anneal(2);
end
if b_pribl_anneal_new(3)>f_max || b_pribl_anneal_new(3)<f_min;
b_pribl_anneal_new(3)=b_pribl_anneal(3);
end
%use the residual function to calculate the probability of a jump
%we calculate the residual function for the candidate for the solution
sigma_new=0.5*sum(((b_pribl_anneal_new(1).*sin(2*pi*b_pribl_anneal_new(3).*x))./(b_pribl_anneal_new(2)^2+x.^2)-y).^2);
%choose a solution for the next iteration
if sigma_new<sigma(i)
b_pribl_anneal=b_pribl_anneal_new;
else
%the difference in the "energies" of the candidates for the solution
d_sigma=sigma_new-sigma(i);
%jump probability
P=exp(-d_sigma/t(i));
end
if P>=rand();
b_pribl_anneal=b_pribl_anneal_new;
end
end
%Interpolation of experimental points using the cubic spline of defect 1
%set of points to interpolate
x_cubic=x;
lambda_1=0;
lambda_n=0;
%calculation of cubic spline of defect 1
[y_cub, coeff]=cub_def(x, y, x_cubic, lambda_1, lambda_n);
%plot of calculated data
figure;
plot(x_cubic, y_cub, 'LineWidth', 3, 'Color', 'g');
hold on;
plot(x, y, 'LineStyle', 'none', 'Marker', '*');
grid on;
%First order differential equation solution
b1=b_pribl_anneal_new(1);
b2=b_pribl_anneal_new(2);
b3=b_pribl_anneal_new(3);
V0=0;
dx=x(2)-x(1);
%improved Euler method for a discrete set of points
V_num_Eu=zeros(size(x));
%initial values
V_num_Eu(1)=V0;
%vector with discrete values
a=x.*sqrt(b1*sin(2*pi*b3*x)/(b2^2+x.^2));
for i=2:length(x)
V_num_Eu(i)=a(i)*dx+V_num_Eu(i-1);
end
%Runge-Kutta method for regression dependence
V_num_RK=zeros(size(x));
%initial values
V_num_RK(1)=V0;
for i=2:length(x)
K1=a(i-1);
K2=(x(i-1)+dx/2).*sqrt(b1*sin(2*pi*b3*x(i-1)+dx/2)/(b2^2+(x(i-1)+dx/2).^2));
K3=K2;
K4=a(i);
V_num_RK(i)=V_num_RK(i-1)+dx/6*(K1+2*K2+2*K3+K4);
end
%Runge-Kutta method for interpolation spline
addpath(pwd);
[a_spline, a_coeff_spline]=cub_def(x, a, x, 0, 0);
for i=2:length(x)
K1=a(i-1);
K2=(x(i-1)+dx/2).*sqrt(a_coeff_spline(i-1, 1)*(dx/2)^3+a_coeff_spline(i-1, 2)*(dx/2)^2+...
a_coeff_spline(i-1, 3)*(dx/2)+a_coeff_spline(i-1, 4));
K3=K2;
K4=a(i);
V_num_RK_sp(i)=V_num_RK(i-1)+dx/6*(K1+2*K2+2*K3+K4);
end
figure
plot(x, V_num_Eu, 'Color', 'y', x, V_num_RK, 'Color', 'g', x, V_num_RK_sp, 'Color', 'r');
hold on
function y_cubic=cub_def(x, y, x_cubic, lambda_1, lambda_n);
%number of experimental points;
n=length(x);
%vectors variables for spline coefficients
a=zeros(n-1,1);
b=a;
c=a;
d=y(1:end-1);
h=diff(x);
delta=diff(y);%finite differences+transposition
phi=delta./h;
theta=h(1:end-1)./h(2:end);
%first and dummy (selected) b_n
b(1)=lambda_1/2;
b(end+1)=lambda_n/2;
%vector of free terms for the matrix search system of all b
%all the same except the first and last
right_side =3./h(2:end).*diff(phi);
right_side(1)=right_side(1)-theta(1)*lambda_1/2;
right_side(end)=right_side(end)-lambda_n/2;
%tridiagonal matrix
%for convenience, there are 3 columns in 1 variable
m=zeros(n-2,3);
%main diagonal of the matrix
m(:,2)=2*(1+theta);
%side diagonals
m(1:end-2,1)=theta(2:end-1);
m(1:end,3)=ones(n-2,1);
M=spdiags(m, [-1 0 1], n-2, n-2);
%calculation of all 's except those already set
b(2:end-1)=M\right_side;
%calculation of all c's except those already set
c=phi-h/3.*(2*b(1:end-1)+b(2:end));
%calculation of all a's
a=1./(h.^2).*(phi-c)-b(1:end-1)./h;
%all values for the spline
y_cubic=zeros(size(x_cubic));
%pass through all intervals [x(i), x(i+1)] between experimental points
%number of points to build a spline
l=length(x_cubic);
y_cubic=zeros(size(x_cubic));
for i=1:n-1
%pass through all x_cube to see if x_cube(j) is in the interval [x(i), x(i+1)]
for j=1:l
if x_cubic(j)>=x(i) && x_cubic(j)<x(i+1)
y_cubic(j)=a(i)*(x_cubic(j)-x(i))^3+b(i)*(x_cubic(j)-x(i))^2+...
c(i)*(x_cubic(j)-x(i))+d(i);
end
end
end
%for the last point separately
if x_cubic(end)==x(end);
y_cubic(end)==y(end);
end
end

Réponses (1)

Les Beckham
Les Beckham le 11 Juil 2022
You are defining this function as a function that takes 5 input arguments and returns one output argument.
function y_cubic=cub_def(x, y, x_cubic, lambda_1, lambda_n);
However, when you call it you are asking for two output arguments.
[y_cub, coeff]=cub_def(x, y, x_cubic, lambda_1, lambda_n);
If you need the function to return two arguments, you must define it that way.
  2 commentaires
Dmitry Surov
Dmitry Surov le 12 Juil 2022
Modifié(e) : Dmitry Surov le 12 Juil 2022
if i define it as
function [y_cub, coeff] =cub_def(x, y, x_cubic, lambda_1, lambda_n);
it still doesn't work..
Les Beckham
Les Beckham le 12 Juil 2022
What do you mean by "it still doesn't work"? Do you get an error message? If so, please post the complete error message (everything in red).
Also, I don't see the variable 'coeff' anywhere in the cub_def function so using that variable as a return argument won't work.
Another thing I just noticed is that
addpath(pwd)
is completely unnecessary. Matlab always searches in the current directory first.

Connectez-vous pour commenter.

Catégories

En savoir plus sur Interpolation dans Help Center et File Exchange

Produits


Version

R2019b

Community Treasure Hunt

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

Start Hunting!

Translated by