bit string ga tool

I want way to represent the value of variable (X) from (1 to 16) or (1 to 32) by using Bit string type gatool (matlab optimization tool)with out repeating any value of X ?

I used this way x=x(1)+x(2)*2+x(3)*4+x(4)*8+x(5)*16 but the matlab remained running with repeating the value of x for example x=4,x=7,x=1,x=1,x=5,x=7,x=13,....etc. with out stopping

[Walter Roberson]

How does this differ from your previous question http://www.mathworks.co.uk/matlabcentral/answers/57616-genetic-algorithm-tool-bit-string ?

[mohammed sportman]

in this way i detected the bit string type

[Walter Roberson]

I still do not understand how it is different. It looks to me to be just the same as the part at the end of your previous question where you remarked about it running endlessly.

Answer 1

[Walter Roberson]

If you want each value to be used exactly once, you should be using a "for" loop or a vectorized computation, not ga(). ga() does adaptive fitting on the value of each variable. If the random computations are leading towards the tendency that x(3) is (say) 0, then x(3) = 0 will be favored in the computations over x(3) = 1. ga() will not systematically try all the possibilities.

Is there a reason why you are not using IntCon and bounds of [1 16] or [1 32] ?

http://www.mathworks.com/help/gads/mixed-integer-optimization.html

[mohammed sportman]

I have a Master project and needed compared to my work in the first two using MATLAB directives have been used(for loop) and the second must be using a genetic algorithm . when i using Double type (integer) the ga tool using single number(0.00 ... point)i do not need that. i want just real number so i used Bit string

[Walter Roberson]

It does not sound to me as if you have tried using intCon

[mohammed sportman]

i tried to use intCon [x,fval,exitflag] = ga(fitnessfcn,nvars,A,b,[],[],...lb,ub,nonlcon,IntCon,options) but my program have one variable X with out A or b or lb .... when i remove them(A or b ...etc) i get an error message . i want to put my program in this equation intCon. can you help me to how i can do that ????

 code                       
    function g = myfitness2(X)
    LAMDA=X(1)+X(2)*2+X(3)*4+X(4)*8+X(5)*16

    TYPE=1;
    ITER=1000;
    R1=Routing3a(TYPE,LAMDA,ITER);
    g=-R1
     end
  -----------

i just want to put the value of LAMDA or(X) as intCon and the value of it between [1 to 16] or [1 to 32]

[Walter Roberson]

Please reformat your comment.

http://www.mathworks.com/matlabcentral/answers/13205-tutorial-how-to-format-your-question-with-markup

[mohammed sportman]

i reformat it

[Walter Roberson]

nvars = 1;
[x,fval,exitflag] = ga(@myfitness3, nvars, [], [], [], [], 1, 16, nonlcon, 1, options);

function g = myfitness3(x)
  LAMBDA = X;
  TYPE = 1;
  ITER = 1000;
  R1 = Routing3a(TYPE, LAMBDA, ITER);
  g = -R1;
end

[mohammed sportman]

When i use the code Who sent to me i get an error like that

Function definitions are not permitted in this context.

and when i using it at command line

nvars = 1;

[x,fval,exitflag] = ga(@myfitness3, nvars, [], [], [], [], 1, 16, nonlcon, 1, options);

i get this error Undefined function or variable 'nonlcon' then i define nonlcon=1

another error appear

Undefined function or variable 'options'*.

when i correct it this error appear

Too many input arguments.

where is the error?????

[Walter Roberson]

function answers56718
  nvars = 1;
  [x,fval,exitflag] = ga(@myfitness3, nvars, [], [], [], [], 1, 16, [], 1, []);
end

function g = myfitness3(x)
  LAMBDA = X;
  TYPE = 1;
  ITER = 1000;
  R1 = Routing3a(TYPE, LAMBDA, ITER);
  g = -R1;
end

[mohammed sportman]

I get this error

Too many input arguments.

[Walter Roberson]

Which routine does it complain about?

It would not hurt to get rid of the final [] making it

[x,fval,exitflag] = ga(@myfitness3, nvars, [], [], [], [], 1, 16, [], 1);

[mohammed sportman]

what you mean please ? i can not understand .

I get this error

Error using ==> ga at 250 Tenth input argument must be a valid structure created with GAOPTIMSET.

[Walter Roberson]

Which MATLAB version are you using? It appears yours might not be recent enough to support integer constraints.

[Walter Roberson]

Try this:

function answers56718
  nvars = 1;
  [x,fval,exitflag] = ga(@myfitness3, nvars, [], [], [], [], 0.5, 16.5);
end

function g = myfitness3(X)
  LAMBDA = round(X);
  TYPE = 1;
  ITER = 1000;
  R1 = Routing3a(TYPE, LAMBDA, ITER);
  g = -R1;
end

[mohammed sportman]

version 7.12.0.635 R(2011a)

[mohammed sportman]

it is about 40 minutes and matalb program still running with out any result

busy

[mohammed sportman]

4 hours

[mohammed sportman]

after 4 hours i get this message

Optimization terminated: average change in the fitness value less than options.TolFun

with out round(x) i use this function

[x,fval,exitflag] = ga(@myfitness33, nvars, [], [], [], [], 1, 16);

function g = myfitness33(X)

    LAMDA = X;
    TYPE = 1;
    ITER = 1000;
    R1 = Routing3a(TYPE, LAMDA, ITER);
    g = -R1;
  end

[Walter Roberson]

Yes, that is a normal termination. Remember that ga does not know how many minima there are, so it is going to keep searching hoping to find a combination that provides a better result. Eventually the search will get to the point where searching everywhere "near" the minimum value that it found is not going to find any change in value. ga is not going to know that that means that value is the minima, it is just going to know that the changes it can detect are very very small. Eventually it will run out of steam and when it does, that is the message it would give. It is completely normal for ga.

[mohammed sportman]

I understand what you mean.But how do I find the solution to the problem I told you in the beginning using the method I find the right solution, but after a long period and the second method find solutions quickly, but it will not stop and, of course, repeat (all using Ga)

[mohammed sportman]

first one

function g =myfitness2(X)
X
s=0;
if (X(1)==1) 
    s=s+1;
end 
if (X(2)==1) 
    s=s+2;
end     
if (X(3)==1) 
    s=s+4;
end  
if (X(4)==1) 
    s=s+8;
end 

LAMDA=s;
TYPE=1;
ITER=1000;
R1=Routing3a(TYPE,LAMDA,ITER);
g=-R1;
 end

[mohammed sportman]

second way

function g =myfitness2(X)
LAMDA=X(1)+X(2)*2+X(3)*4+X(4)*8+X(5)*16
TYPE=1;
ITER=1000;
R1=Routing3a(TYPE,LAMDA,ITER);
g=-R1
 end

[Walter Roberson]

The way to find the solution in minimal time is not to use ga() .

[mohammed sportman]

if i want to use the genetic algorithm in my project what i say (the ga don't support my project)

my super visor asked me to use the ga.

that is problem for me.

[Walter Roberson]

Then you report back to your supervisor that Yes, you can use ga, but that it is very very time consuming and that the results it produces are not better and cannot be better than a very short "for" loop, and then ask how they want you to proceed. If they tell you to go ahead with it anyhow, then you deliver the code that takes 40+ hours for something that should be a fraction of a second.

Now I notice that your Routing3a routine has an "ITER" parameter. What is the purpose of that? Is there some randomization going on in Routing3a? If Routing3a is doing some kind of searching, then it is that searching that you should be automating through ga.

[mohammed sportman]

Thank you Walter Roberson for your valuable with me constantly.Draft works at the expense of the number of requests accepted and rejected for optical networks.The possibility of increasing the number of services using optical fiber.Where lambda is the channel inside the optical fiber and increase the number of it means to increase the number of requests accepted. I have a major program calls paths function to find all shortest paths.We have completed the programming using MATLAB directives. Supervisors asked me to improve it using genetic function because the main business of GA in any function is optimization . And my time to complete the project narrow somewhat (few) so I tried to use tools function genetic to improvement project as a whole and by comparing running and find lamda appropriate to create a state saturation 100% where all requests are accepted (no rejection) in terms of time, but I was surprised where that time the situation normal less much of the use of genetic function. There is a part of my project is to find all paths if there was a ready function used in the genetic algorithm please help me out. And if you can help me in any part of the project, please tell me so and I will send detail required.... with my respect and appreciation

Routing3a it the main function in project.

ITER number of iteration .The more increase the accuracy of the results, but does not have many teams for my project