partially mapped crossover function for tsp

Hi,

I have 2 random parents data example,

P1: 4 8 7 | 3 6 5 | 1 10 9 2

P2: 3 1 4 | 2 7 9 | 10 8 6 5

I need to perform partially mapped crossover function to find a new child.

C1: 4 8 6 | 2 7 9 | 1 10 5 3

C2: 2 1 4 | 3 6 5 | 10 8 7 9

help me in coding this

Thanks in Advance..

Answer 1

[Walter Roberson]

Use a CrossOverFcn of 'crossovertwopoint' https://www.mathworks.com/help/gads/genetic-algorithm-options.html#f7820

[msr16]

for tsp path generation 2 point cross over will give the repetation of cities.......

[Walter Roberson]

Okay, then write your own custom crossover function that follows whatever cross-over principles you want.

Answer 2

[Arunkumar Gopu]

Yes, i do understand that you shoud not have repeated values when working with TSP problem in you sequence of cities. You can do PMX for such kind of problem and the code goes here.

function child= crossover(a,b)

    crossoverrate=0.3;

    [~,sizea]=size(a);
    [~,sizeb]=size(b);

    child1=zeros(1,sizea);
    child2=zeros(1,sizea);

    if sizea~=sizeb
        print("a and b are not equal cant do crossover");
    end


    for i=1:100
        %this line will generate 2 random values
        lo=randi(sizea);
        rate=ceil(sizea*crossoverrate);
        up=lo+rate;
        if up>sizea
            continue;
        else
            break;
        end
    end

    
    for i=lo:up
        child1(i)=b(i);
        child2(i)=a(i);
    end
    
    
    % child1
    for i=lo:up
        d=0;
        for j=lo:up
            if(a(i)==child1(j))
                d=d+1;
                break;
            end           
        end

        if d==0
         crossover1(i,i);              
        end    
    end

    function  crossover1(n,i)
       for k=1:sizea

                if child1(n)==a(k)
                   if child1(k)==0
                      child1(k)=a(i);
                      break;
                   else
                       crossover1(k,i)

                   end 

                end
       end
     end

  
    for q=1:sizea
        if child1(q)==0
            child1(q)=a(q);
        end       

    end
    % child2
    for i=lo:up
        d=0;
        for j=lo:up
            if(b(i)==child2(j))
                d=d+1;
                break;
            end           
        end

        if d==0
         crossover2(i,i);              
        end    
    end

    function  crossover2(n,i)
       for k=1:sizeb

                if child2(n)==b(k)
                   if child2(k)==0
                      child2(k)=b(i);
                      break;
                   else
                       crossover2(k,i)

                   end 

                end
       end
     end

  
    for q=1:sizeb
        if child2(q)==0
            child2(q)=b(q);
        end       
    end
    
    %return the values here
    child=[child1,
        child2];
end

[Walter Roberson]

More comments would help.

[Walter Roberson]

Even more comments than that would help. For example what is the purpose of your two extra functions? What is their "interface contract" (inputs requirements, output types) ? What algorithm is being used?