Galapagos Optimisation Logic

Hi all,

I’m presented with an interesting problem which I hoping some of you can share some thoughts on.

I would like to minimise the maximum displacement of this orthotic brace whilst keeping the weight at an absolute minimum. This is achieved by selectively allocating a varying percentage of carbon fibre to the shell. Carbon fibre can be added up to 30%, however, this increases weight.

My current approach is to randomly pick vertices from the mesh and allocate carbon fibre to the surrounding area by a given radius. The Galapagos optimisation is simply defined as:

Fitness = Max Displacement * Total Weight (both are normalised)
Genome = [Random Seed, Number of material points (spheres)]

The point cloud in the picture below represents Young’s modulus of the object as varied by the addition of carbon fibres.

One of the first points would be to improve how the points are selected, I understand picking them randomly is less than ideal for Galapagos given the discontinuity. Exploring the mesh perhaps using a UV value seems better suited however it’s not clear to me how to make use of the mesh parameters?


Upload file

I’d love to but I’m working under NDA’s, further you’d need a Karamba licence to run the definition.

Than I wouldn’t upload a picture. Karamba license isn‘t a problem for everyone. Create an example and upload it, so we don‘t have to start from scratch

About the selection. If I understand you right, you run the analysis after selecting the area for the reinforcement. How about running the analysis. Select the area with most utilization and add there carbon fiber. Loop this till you get a nice distribution.