Why it gives error if we set lb=[0 0 0 0], ub=[10 10 pi pi] instead of lb=0 and ub=pi?

I download an algorithm from Mathworks site called "Student Psycology Based Optimization (SPBO) Algorithm" located on URL: https://www.mathworks.com/matlabcentral/fileexchange/80991-student-psycology-based-optimization-spbo-algorithm?s_tid=FX_rc1_behav

My fitness function is as:

function out=myfitness(b)
  u = [1 2 3 4];
  out1 = u-b;             %needs R2016b or later
  out = sum(out1.^2,2);
end

When I type the following on command window, it runs

[Best_fitness,Best_student,Convergence_curve]=SPBO(20,100,pi,0,4,@myfitness)

But when I type the following on command window:

[Best_fitness,Best_student,Convergence_curve]=SPBO(20,100,[10 10 pi pi],[0 0 0 0],4,@myfitness)

it gives the following error:

Assignment has more non-singleton rhs dimensions than non-singleton subscripts
Error in SPBO (line 119)
                     par1(dw,do)=mini+(rand*(maxi-mini));                    % Equation (4) 

How to handle this?

Answer 1

[Walter Roberson]

It is a bug in the code.

The code also regularly uses ans as a variable name to be stored into and recalled from, which makes the code fragile, since whether any particular construct sets ans or not depends upon the MATLAB release, and people reading the code (and repairing it) should not be expected to know all of the rules.

[Walter Roberson]

par1(dw,do)=mini+(rand*(maxi(do)-mini(do)));                    % Equation (4)

But first you need to insert something near the top,

%expand mini and maxi out to length same as number of variables
mini(end:variable) = mini(end);
maxi(end:variable) = maxi(end);

Answer 2

[Cris LaPierre]

From the description of SPBO

% variable = number of your variables
...
% mini=[mini1,mini2,...,minin] where mini is the lower bound of variable n
% maxi=[maxi1,maxi2,...,maxin] where maxi is the upper bound of variable n

The syntax for calling SPBO is

% To run SPBO: [Best_fitness,Best_student,Convergence_curve]=SPBO(student,Max_iteration,mini,maxi,variable,fobj)

You appear to have flipped your mini and maxi. You should first input the mini values, then the maxi. I would expect mini should be less than maxi. Try this calling syntax.

[Best_fitness,Best_student,Convergence_curve]=SPBO(20,100,[0 0 0 0],[10 10 pi pi],4,@myfitness)

Just a note that I don't use this toolbox, and my answer is untested. Just seems likely to cause an error is it probably violates an unstated/untested assumption the code author made.

[Sadiq Akbar]

Dear Cris LaPierre thanks for your response. No its not like that. If you download it and see inside main.m or even the definition line of SPBO, its the one I am using. i.e.

>>[Best_fitness,Best_student,Convergence_curve]=SPBO(student,Max_iteration,maxi,mini,variable,fobj)

[Sadiq Akbar]

Dear Walter Roberson already I have put you in problem. But since you commented, let me ask you, how can we remove the bug?

[Cris LaPierre]

Ah ok. In that case, I defer to Walter's response.

[Sadiq Akbar]

Thanks for your response dear Cris LaPierre