Arrays have incompatible sizes for this operation.

Hello guys , I need help to fix my code . "It is giving me an error of

Error in untitled8 (line 19)"

A = diag(alpha, -1) + diag(beta) + diag(gamma, 1);

The final plot should show an ellipse with 8 points . Generally I want to create a parametric cubic spline with boundary conditions to this ellipse.

Any suggestions will be highly appreciated .

Here is my code (script below)

% Number of points on the ellipse
N = 8;
% Parametric representation of the ellipse
theta = linspace(0, 2*pi, N);
x = cos(theta);
y = sqrt(2)*sin(theta);
% Periodic parametric cubic spline interpolation
t = linspace(0, 2*pi, 1000);
ti = linspace(0, 2*pi, N);
% Build the tridiagonal system for solving the cubic spline coefficients
h = diff(ti);
alpha = 1/6 * h;
beta = (h(1:end-1) + h(2:end)) / 3;
gamma = 1/6 * h(2:end);
A = diag(alpha, -1) + diag(beta) + diag(gamma, 1);

Arrays have incompatible sizes for this operation.

A(1, N) = 1/6 * h(end);
A(N, 1) = 1/6 * h(1);
% Compute second derivatives (spline coefficients)
b = 6 * diff(diff([x, x(1)])./h);
c = A\b';
% Evaluate the cubic spline at the specified points
spline_x = zeros(1, length(t));
spline_y = zeros(1, length(t));
for i = 1:N
    indices = (t >= ti(i)) & (t < ti(i+1) | (i == N && t == 2*pi));
    dt = t(indices) - ti(i);
    spline_x(indices) = c(i) / 6 * (ti(i+1) - dt).^3 + c(i+1) / 6 * dt.^3 + (x(i) - c(i) * h(i)^2 / 6) * (ti(i+1) - dt) + (x(i+1) - c(i+1) * h(i)^2 / 6) * dt;
    spline_y(indices) = c(i) / 6 * (ti(i+1) - dt).^3 + c(i+1) / 6 * dt.^3 + (y(i) - c(i) * h(i)^2 / 6) * (ti(i+1) - dt) + (y(i+1) - c(i+1) * h(i)^2 / 6) * dt;
end
% Plot the results
figure;
plot(x, y, 'o', 'MarkerSize', 10, 'DisplayName', 'Ellipse Points');
hold on;
plot(spline_x, spline_y, 'LineWidth', 2, 'DisplayName', 'Periodic Cubic Spline Interpolation');
axis equal;
legend;
title('Periodic Parametric Cubic Spline Interpolation on Ellipse');
xlabel('x');
ylabel('y');
grid on;

Answer 1

[Walter Roberson]

% Number of points on the ellipse
N = 8;
% Parametric representation of the ellipse
theta = linspace(0, 2*pi, N);
x = cos(theta);
y = sqrt(2)*sin(theta);
% Periodic parametric cubic spline interpolation
t = linspace(0, 2*pi, 1000);
ti = linspace(0, 2*pi, N);
% Build the tridiagonal system for solving the cubic spline coefficients
h = diff(ti);
alpha = 1/6 * h;
beta = (h(1:end-1) + h(2:end)) / 3;
gamma = 1/6 * h(2:end);
whos alpha beta gamma

  Name       Size            Bytes  Class     Attributes

  alpha      1x7                56  double              
  beta       1x6                48  double              
  gamma      1x6                48  double              

A = diag(alpha, -1) + diag(beta) + diag(gamma, 1);

Arrays have incompatible sizes for this operation.

alpha is 1 x 7. To put it one off the diagonal requires that the array be 8 x 8.

beta is 1 x 6. To put it on the main diagonal requries that the array be 6 x 6.

gamm is 1 x 6. To put it one of the diagonal requires that the array be 7 x 7.

So you are trying to add together an 8 x 8 with a 6 x 6 and a 7 x 7.

[Panda05]

Thank you so much . so that need to be changed