Crossover function in gene expression programming

Hi, I want to write a crossover function for GEP, right now i am stuck at the part replacing the subtree into the designated crossover point, I have tried chatgpt numerous times but it keeps on assign old_tree to new_tree which makes the crossover function returns old_tree, here is my crossover code:

function new_tree = set_subtree(old_tree, crossover_point, subtree)

if crossover_point == 1

new_tree = subtree;

return

end

queue = {old_tree};

idx = 1;

while ~isempty(queue)

node = queue{1};

queue = queue(2:end);

if isfield(node, 'children')

for i = 1:length(node.children)

idx = idx + 1;

if idx == crossover_point

node.children{i} = subtree;

new_tree = old_tree;

return

end

queue{end+1} = node.children{i};

end

end

end

error('Crossover point not found in expression tree');

end

Thank you for your time.

Answer 1

[Rishav]

Hi LONG,

To fix this issue, you can create a copy of old_tree before modifying it, and then assign the modified tree to the new copy.

Here is an updated code by changing the set_subtree function if index equals crossover point:

if idx == crossover_point
      % Create a copy of the old tree
      new_tree = old_tree;
      % Replace the subtree at the crossover point
      new_tree.children{i} = subtree;
      return
end

In this updated version, new_tree is initialized to a copy of old_tree before the subtree is replaced at the designated crossover point. This ensures that new_tree is not simply a reference to old_tree, but a new copy that includes the modified subtree.

Regards,

Rishav Saha