How to formulate GA with plenty variables (e.g >24000 variables)

Hi everyone.

Please what is the best way to formulate a GA using the GA toolbox in MATLAB. Most of the examples I have seen are limited to at most 10 variables that are collected in an array X.

In my problem, variables e.g ON status of a unit changes with respect to the time or period of 24 hours. Which means for every UNIT, it has 24 variables just indicating the ON status of the unit throughout the 24 hour period.

I have multiple units in my problem which means variables for ON status alone is already 24 * Number of Units (1000).

Please what is the best way to formulate this problem for solving using GA toolbox?

[praveen kumar]

ga tool box cannot be used u can write your own code

[praveen kumar]

Instead of using binary representation for unit commitment you can represent ON/OFF status as integer coded this reduces the memory required for ON/OFF status of Unit commitment..Here his the code for UC problem with 10 unit system 24 hrs sheduling period

clear all;
close all;
clc;
global data n h Pd Pmax CC CCC pop_size YY
pop_size=1;
n=10;
h=24;
cycles=5;
Penalty=100;
data=[455 150 1000 16.19 0.00048 8 8 4500 9000   5  8
      455 150 970  17.26 0.00031 8 8 5000 10000  5  8
      130 20  700  16.60 0.00200 5 5 550  1100   4 -5
      130 20  680  16.50 0.00211 5 5 560  1120   4 -5
      162 25 450   19.70 0.00398 6 6 900  1800   4 -6
       80 20 370   22.26 0.00712 3 3 170  340    2 -3
       85 25 480   27.74 0.00079 3 3 260  520    2 -3
       55 10 660   25.92 0.00413 1 1 30   60     0 -1
       55 10 665   27.27 0.00222 1 1 30   60     0 -1
       55 10 670   27.79 0.00173 1 1 30   60     0 -1];
 
Pd=[700 750 850 950 1000 1100 1150 1200 1300 1400 1450 1500 1400 1300 1200 1050 1000 1100 1200 1400 1300 1100 900 800];
SR=0.1.*Pd;
Pmax=data(:,1);
Pmin=data(:,2);
a=data(:,3);
b=data(:,4);
c=data(:,5);
minup=data(:,6);
mindown=data(:,7);
hotstartcost=data(:,8);
coldstartcost=data(:,9);
coldstarthrs=data(:,10);
initialstatus=data(:,11); 
  
 CC=zeros(n,cycles,pop_size);
% base unit
for k=1:pop_size
    for i=1:2
CC(i,1,k)=24;
    end
end
%initialization  on/off shedule based on  load curve
for k=1:pop_size
for i=3:n
     temp1=0;temp2=0;
    for j=1:cycles   
                  if(initialstatus(i)<0 && j==1 )
                 temp=max(0,(minup(i)+initialstatus(i)));
                 CC(i,j,k)=-round(unifrnd(temp,h));
                 
             end
             if(initialstatus(i)>0 && j==1)
                 temp=max(0,(minup(i)-initialstatus(i)));
                 CC(i,j,k)=round(unifrnd(temp,h));
                    temp1=sum(abs(CC(i,j,k)));
                    temp2=h-temp1;
                    
             end
          
          
              if(j>1&& CC(i,j-1,k)<0)
                  if(temp2>minup(i)) 
                     CC(i,j,k)=round(unifrnd(minup(i),temp2));
                 
                 end
               if(temp2<=minup(i))
               CC(i,j,k)=temp2;
              
               end
            end
           
           if(j>1&& CC(i,j-1,k)>0 )
               if(temp2>minup(i)) 
               CC(i,j,k)=-round(unifrnd(minup(i),temp2));
               end
               if(temp2<=minup(i)) 
               CC(i,j,k)=-(temp2);
               
               end
           end  
           temp1=temp1+sum(abs(CC(i,j,k)));
           temp2=h-temp1; 
         
          end
               
           end    
  
end
CC

CC = 10×5

24 0 0 0 0 24 0 0 0 0 -2 11 -5 6 0 -7 10 -7 0 0 -13 9 -2 0 0 -1 4 -8 4 -7 -20 4 0 0 0 -19 2 -2 1 0 0 0 0 0 0 0 0 0 0 0

Answer 1

[Alan Weiss]

It seems that you have thousands of binary variables. If so, then indeed ga is not an effective solver for your problem.

You would likely do better using intlinprog as a solver. You need to formulate your problem as mixed-integer linear programming, but, if you look at the examples, you will see that a wide variety of problems can be formulated that way.

Good luck,

Alan Weiss

MATLAB mathematical toolbox documentation