Same code in different system giving different output

I’ve written a code that collects data from an Excel file, then enters a while loop for iteration using fminsearch along with two custom functions I’ve developed. The loop exits based on a specific breaking condition. All my files are stored in a OneDrive folder, which I can access both from my desktop and my laptop.

However, I’m experiencing an issue: When I run the code on my desktop, it completes 114 iterations, while on my laptop, it runs for 314 iterations. Additionally, the output values differ between the two systems. Since the code file is located in OneDrive, I’m using the exact same code (without copying or making separate files). The only difference in the code is how it reads the Excel file—I modify the file path depending on whether I’m working on my desktop or laptop, and I handle this through commented-out lines. I am using the same version of matlab in both system.

Could anyone kindly provide insight into why this is happening? I would also appreciate any suggestions on how I can ensure consistent results across both systems.

[Mike Croucher]

Is it possible to show the code and data?

[H M Dilshad Alam]

clear all


%dataset for laptop
data = xlsread("locationfromlaptop\1.xlsx");

%dataset for Lab pc
%data = xlsread("locationfromdesktop\1.xlsx");
x = data(:, 1:5);
y = data(:, 6);   

beta = [719.9786; -80.0245; -80.8986; 20.72844; -0.54265; -211.659; -35.3204; 
        16.67953; -0.96645; 0.104964; 31.0675; 271.1223; 4.225911; -0.09841; 
        69.22954; -0.91565; 0.002612; -7.39912; 3.05E-05; 0.114144; 74.11468];

tolerance = 1e-5;
max_iterations = 1000; 
iteration = 0;
fvalthetas = [];
fvalbetas = [];


theta0 = zeros(1, 21); 

while iteration < max_iterations
    iteration = iteration + 1;
    
    options = optimset('Display', 'iter');
    [theta, fvaltheta] = fminsearch(@(theta) computeSumNew(theta, beta, x, y), theta0, options);
    disp(fvaltheta)
    fvalthetas = [fvalthetas; fvaltheta];
    
    [beta_new, fvalbeta] = fminsearch(@(beta) weightedSumNew(beta, theta, x, y), beta, options);
    disp(fvalbeta)
    fvalbetas = [fvalbetas, fvalbeta];

    disp(iteration)
    if max(abs(beta_new - beta)) < tolerance
        break;
    end
    
    beta = beta_new;
end

disp('Converged Beta:');
disp(beta);
disp('Converged Theta:');
disp(theta);

The functions used in the codes are:

function result = computeSumNew(theta, beta, x, y)
X = [ones(size(x, 1), 1), x, x(:, 1).^2, x(:, 1).*x(:, 2), x(:, 1).*x(:, 3), x(:, 1).*x(:, 4), x(:, 1).*x(:, 5), ...
                  x(:, 2).^2, x(:, 2).*x(:, 3), x(:, 2).*x(:, 4), x(:, 2).*x(:, 5), ...
                  x(:, 3).^2, x(:, 3).*x(:, 4), x(:, 3).*x(:, 5), ...
                  x(:, 4).^2, x(:, 4).*x(:, 5), ...
                  x(:, 5).^2];

output = X * beta;
theta_term = X * theta';

sum_result = 0;
for i = 1:length(y)
    squared_diff = (y(i) - output(i)).^2;
    lggd = log(squared_diff);
    term1 = lggd - theta_term(i);
    term = term1^2;
    sum_result = sum_result + term;
end
result = sum_result;
end

the other one

function result = weightedSumNew(beta, theta, x, y)

X = [ones(size(x, 1), 1), x, x(:, 1).^2, x(:, 1).*x(:, 2), x(:, 1).*x(:, 3), x(:, 1).*x(:, 4), x(:, 1).*x(:, 5), ...
                  x(:, 2).^2, x(:, 2).*x(:, 3), x(:, 2).*x(:, 4), x(:, 2).*x(:, 5), ...
                  x(:, 3).^2, x(:, 3).*x(:, 4), x(:, 3).*x(:, 5), ...
                  x(:, 4).^2, x(:, 4).*x(:, 5), ...
                  x(:, 5).^2];


output = X * beta;

sum_result = 0;
for i = 1:length(y)
    theta_term = X(i, :) * theta';
    wi = exp(-theta_term);
    squared_diff = (y(i) - output(i)).^2;
    term = wi * squared_diff;
    sum_result = sum_result + term;
end
result = sum_result;
end

The data is a excel sheet with first 5 coloums showing 5 variables of x and the 6th coloum is the y value. There are total 50 data points.

[Walter Roberson]

Can you attach the data?