Question Regarding MultiObjective Optimization - gamultiobj

Hi. I want to create a model for multi objective optimization.

I have 300 items and three variables for each. Objective is the a function of using 300 items * 3 variables. So I don't want to write 900 seperate variables in the function, if possible. Thanks a lot in advance.

My code is as follows. Also I receive the following error.

error message

++

Not enough input arguments.

Error in globalfun (line 6)

y(1) = sum((1-x(1,1)*0.99)+(1-x(1,2)*0.95)+(1-x(1,3)*0.90));

Error in MCICv2 (line 26)

[x,fval] = gamultiobj(globalfun,numberOfVariables);

++

Script part

++

clc;clear all;close all;

global leadtime

data = readmatrix('Data.xlsx');

data(:,1) = [1:301];

ProductCodes = data(:,1);

LeadTime = data(:,4); %This is the value I use for Multi-Objective Optimization

[widthdata, heightdata]= size(data);

numberOfVariables = i*3;

[x,fval] = gamultiobj(globalfun,numberOfVariables);

++

Function Part

+++

function y = globalfun(x)

global LeadTime

y(1) = sum((1-x(:,1)*0.99)+(1-x(:,2)*0.95)+(1-x(:,3)*0.90)); %minimize the sum of all 301 items

y(2) = (x(:,1)*norminv(0.99))+(x(:,2)*norminv(0.95))+(x(:,3)*norminv(0.90))*sqrt(LeadTime(:)); %minimize the sum of this function for 301 items

end

+++

[Alan Weiss]

Sorry, I don't quite understand what you are trying to do. How many variables do you have in your optimization, I mean variables that you can control, not data? Do you have 3 variables? 300? 900? I understand that you have two objective functions, but I cannot easily tell which variables you can control and which are data that you do not control.

Alan Weiss

MATLAB mathematical toolbox documentation

[Fatih Yigit]

Thanks for your response. Let me explain this way. I have 300 items.

Each item has a parameter called `LeadTime(:)`.

This parameter will be multiplied with X(:,1) & x(:,2) and (x(:,3) as given in the formula.

++

sum((1-x(:,1)*0.99)+(1-x(:,2)*0.95)+(1-x(:,3)*0.90))

++

So all items will have a single value. The goal is to minimize the sum of the function for 300 items.

In short, I'll have 900 variables. 3 variables will affect a single item with a `LeadTime(n)` . n=1,2,3....300

This part is just the beginnig, but I am even struggling in this part. Hope it is more clear now. Thanks for your time again.

[Fatih Yigit]

To be more precise; I attach the script, function, error in the following lines. If I manually assign x e.g [1:301] function works. But not via "gamultiobj".

This is the main script

++

clc;clear all;close all;

global LeadTime; %301 items

data = readmatrix('Data.xlsx');

data(:,1) = [1:301];

ProductCodes = data(:,1);

LeadTime = data(:,4); %This is the value I use for Multi-Objective Optimization

[widthdata, heightdata]= size(data);

result = gamultiobj(globalv6, 903);

++

this is the function

+++

function y = globalv6(x)

global LeadTime

[width, height] = size(LeadTime); %width is 301

j=0;

for i=1:3:(3*width)

j=j+1;

result(j,1) = x(i)+x(i+1)+x(i+2);

result(j,2) = ((1-x(i)*0.99)+(1-x(i+1)*0.95)+(1-x(i+2)*0.90));

end

y(1) = sum(result(:,1));

y(2) = sum(result(:,2));

end

++

error;

Not enough input arguments.

Error in globalv6 (line 8)

result(j,1) = x(i)+x(i+1)+x(i+2);

Error in MCICv3 (line 13)

result = gamultiobj(globalv6, 903);

Answer 1

[Alan Weiss]

I have a hard time understanding you because you use vocabulary that is not standard. You say that you have items. I ask about control variables and data. Is an item a control variable or data?

data(:,1) = [1:301];

I wonder whether this should be transposed:

data(:,1) = [1:301]';

You show your function call as

 result = gamultiobj(globalv6, 903);

Is globalv6 a function handle? I mean, I would expect the function call to be

 result = gamultiobj(@globalv6, 903); % Note the @ sign

But the more important question is, do you have no constraints on x, not even bounds? For example, should your x variables be positive? Less than 1? I think so, so you would want to incorporate bounds

lb = zeros(1,903);
ub = ones(1,903);
result = gamultiobj(@globalv6, 903,[],[],[],[],lb,ub);

Alan Weiss

MATLAB mathematical toolbox documentation

Answer 2

[Fatih Yigit]

Thanks for your help. You are right. Items are actually that have values used as parameters.

I now know that the main problem is not using "@".

My model will have multiple constraints but I just want to start understanding the use use of "gamultiobj" first. Now it is clear.

For the time being mty problem is solved. Again thanks for your help.