BESO for Beams Background

I have been reading about the BESO for beams component quite a bit over the weekend. I am hoping to try and replicate some of the classic optimisation problems such as Michelle cantilevers before expanding onto current live projects. I have read the paper Evolutionary Structural Optimization Revisited by @cp1, which was really helpful in understanding the component a little better but I do have some additional questions:

  1. BESO requires an input of the target mass ratio. When trying to find an optimal or near-optimal solution the user may not know what ratio to use. Is the intention to use a genetic algorithm to vary this to get to an optimal structure?
  2. The paper describes the use of genetic algorithms to improve the output of the component. It was not clear from the paper if this GA is included in the BESO component or it is intended to use a separate GA such as Galapagos within the definition.
  3. Finally would it be possible to share the definition that the paper is based on? This would be really helpful.

Any help would be greatly appreciated.


Hello @alexander.griffiths,
thanks for your interest in the paper. I am glad that you find it helpful!
ad 1.) The target mass is something that needs to be chosen with respect to the maximum displacement or utilization of the resulting structure. The BESO procedure as implemented in Karamba3D does not take account of the strength of the underlying materials.
ad 2.) The GA is not included in the BESO it was rather used to find good values for the input parameters of the algorithm.
ad 3.) I will go through my archive. If I find the definition I will post it here.
– Clemens