Galapagos does not find the right solution


I am trying to optimize the arrangement of bars in a steel truss structure, manipulating the direction of the diagonals and de-/activating the verticals with sliders (contained in the gene pools), see picture below.
My objective is minimum weight with a penalty for high displacements.
Later on, if it works, I would like to add more variables.

It seems that Galapagos only finds solutions that are worse than the one I find manually. How do I help the algorithm with finding the right solution? Is it the right approach for topology optimization? Why does it matter what solution is start with?

Additional question: How is Karamba considering buckling and where can I find the buckling lengths?

Thanks! (123.8 KB)

Hi @adrnna, did you make sure to select minimize in the Galapagos settings, otherwise Galapagos will optimise for the maximum value, which is therefore the worse solution.

Hi @matttam, yes the goal value is set to be minimized, so that’s not the issue… The solution is close to the optimal one in regards to weight but still not what I am looking for.

Hi @adrnna, when i open your script, and I try to optimise it, it does not seem to work. It seems to related to the parameters used for the optimisation, that you only have 2 variables for each parameter. Try adjust the approach and see if it helps.

In regards to the buckling lengths, this can be located with the disassemble Element component: