Hey there! I’m trying to do a multi-objective optimization minimizing the Co2 emissions and the radiance, but it doesn’t seems to work. I did modified the inputs in case that 0 gives a null geometry. Still, this doesn’t work. I would really appreciate some help since I just started with this plugin! thank you so much <3
try3.gh (205.3 KB)
im unable to open your definition, there are missing plugins (see image).
when you say ‘it doesnt seem to work’… can you provide some more information… what isnt working?
I see. Sorry for that, I’m using a plugin called Cardinal LCA for calculating the embeded CO2. But I’ll upload the definition again without the optimization of that since it doesn’t seems to be detected (do you think this could be because it’s a very random plugin?)
The graph of the CO2 doesn’t show but the one of for radiation, it’s very strange as it shows the image. I think maybe it takes a while for my laptop to make the optimization since it’s a very simple and humble set up.
This is a graph for the radiance optimization. The only sliders that I connected are 4: width and lenght of the panel and angle and position of the attractor.
Please i would be very very grateful if you can provide me any opinion in the matter if there’s something wrong since I’m trying to do this for a task in my bachelor degree in architecture.
Thank you so much <3 and here is the definition without the LCA cardinal
try3.gh (209.6 KB)
from what i can see from the screenshots, the fitness values for the second objective (minimise GWP) do not change when the slider values change. When wallacei changes the sliders (the genes) to create a new solution, the change in the slider values should impact and change the fitness values you are optimising… if this doesnt happen, then the graph for that particular fitness objective will not appear, which then causes display issues with the other graphs such as the parallel coordinate plot, as is currently happening in your screenshots. if you disable the second objective (like you did in the grasshopper definition you sent me) and run the simulation for just the first objective, you should see the graphs emerge. to fix this problem, make sure your genes (i.e. sliders) in grasshopper have the “wlc_” prefix in their name, and then use the wallacei genepool shuffler to test if when the genes shuffle, does it change the fitness value, and then debug from there.
With regards to the first fitness objective, there is nothing wrong (technically) with it, however the reason you are seeing a thin long line is because the two generations that the algorithm evolved (gen 0 and gen 1) have a massive difference in how much variation there is within the population that when the charts are remapped, you see a very thin tall line, which is cause by a very flat and wide line (which you cant see in the chart). This is a design problem issue… what i recommend is you stop the simulation after the generation 1 finishes, and then export the solutions that you think are causing this problem, usually there would be one solution that is highly ‘mutated’ and so its fitness value is very different then the rest of the population. You should be able to identify this solution through using the charts in tab 2 of wallacei.
Hope the above helps!
Hello! Thank you so much for your response. I did make it work with the GWP slider and none of them now goes to 0. I can see Wallacei running (in rhino i can see the changing of geometry) but in the wallacei component there is no data showing.
Do you know why this could be happening? I used a small population size since it was a test but, before I put 150 total and happened the same.
Thank you so much again for your help.
nvm it worked!!! im so happy!! i just needed to run the analysis for radiance and daylight