Creating an Optimized Table Support Structure

For project I'm working on I'd like to try my hand at structurally optimizing the frame that would hold up the top of a table (will most likely will be glass). I will be CNC-ing the mold to cast the structure I come up with in concrete, but I'm not sure if my script is accomplishing what I want it to in the most efficient way.

Currently, I'm using millipede to give me the stress of each "beam" in the beam system I've created, and then I've given galapagos control of some sliders that affect the position of the nodes and set it to minimize the average amount of stress (which is given by the millipede FE Solver)

Essentially I'm curious if the way I'm going about this makes sense, particularly the way I have the galapagos evolutionary solver set up, since some of the results I'm getting don't seem to make sense intuitively.