Hello everyone,
Recently, I am trying to use Walacei. But I have a question on the pareto front in Wallacei Analytics. As you can see from the screenshot1, the pareto front solutions are clearly the yellow ones, which are supposed to be equal to the pareto front solutions in tab 3. But, in Wallacei Analytics, when I input the required generation, the results are always the solutions in one generation, not the pareto ones.Any suggestions?
Best,
Green
By the way, I tried to compare the pareto front solutions in tab3 with the pareto front solutions in the last generation. Clearly, they don’t match very well. Pareto front solutions in the last generation are supposed to be part of all the pareto front solutions in tab3.
Please share your file I will look at this. One note, You are asking to see the pareto front of all generations in the UI rather than only the last generations, there is a slight differences due to repetitions of the solutions.
Hi Milad,
The file I use is the example file of Wallacei.
Wallacei X_Example File.gh (307.4 KB)
1- yellow pareto fronts in tab1 are the same in pareto fronts in tab 3 but you need to make sure you show the right ones, tab 1 has the last generations and the entire populations same as the tab 3.
2-wallacei analytics show the pareto front in the selected generations, it does not have an option to show the entire population, the pareto front per generations are as “real” as the pareto front from all population. the pool for analysis is different, not the none dominance calculations.
3- the coordinates are remapped and projected in a domain to be easily visualised and compared. you might have projected them with different remapped values.
Hi Milad,
Thanks a lot. As for your answer 2, I am still a little confused. In the example file, I don’t think the green cubes generated by “Wallacei Analytics_Objective Space” are pareto front solutions. I think they are all the solutions from every generation. And, once I connect “Objective Space” to “Wallacei Analytics_Pareto Front”, the component doesn’t give me the pareto front solutions. However, it gives me all the solutions for one generation. If I have a population size of 100, I will get 100 cubes in “Wallacei Analytics_Pareto Front”, which I don’t think is right. Pareto front solutions should be non-dominated Solutions.
The objective space component does not give you pareto front solutions, the green cubes are all solutions, there is another component that gives the pareto which you have it on the canvas, it gives you the pareto front per generation, you can give the last generation and it gives the pareto front for the last generation
Hi Milad,
Could you tell me the name of the component?
“Wallacei Analytics_Pareto Front” gives the solutions, but I don’t think they are pareto front solutions, the number of which is the same amount of the population size for each generation.
The ‘Wallacei Analytics_Pareto Front’ component allows you to choose a specific generation to analyse and extract the Pareto front, but it doesn’t automatically output this information for you. When you choose a generation to identify the Pareto Front, what this component does is give you the ‘Dominance Rank’ for each solution in the selected generation. It also outputs the mesh cubes for each solution in the given generation… and so all you need to do is dispatch the mesh cubes using the dominance rank, in which a dominance ranking of 0 will give you the Pareto Front. The reason we did this is that this allows users to choose solutions with any dominance rank (not just 0). Of course the Pareto Front are the solutions that most people are interested in (i.e. dominance rank of 0), but this approach also allows for identifying solutions with other dominance ranking. Here below is the process explained (in generation 5, there are 26 pareto front solutions):
Hi Mohammed,
Thanks a lot. I have more questions according to your answer. I found that the the number of pareto front solutions in one generation does not match that in tab 3. As you can see, the number of pareto front solutions in 20th generation is 50 vs 19.
Is this because the component only calculates one generation, while the tab3 calculates all generations?
Correct.
The Pareto Front calculation in Wallacei X Tab 3 calculates the dominance value by analysing ALL the solutions in the entire population (in this case, 5000 solutions) while the Pareto front calculation in the Wallacei analytics components analyses the solutions within a given generation only (in this case, 50 solutions).
In the example you provided, when calculating the dominance value of solutions in generation 20 only, it shows that all the solutions have a dominance value of 0. However, when you calculate the pareto front for all the solutions in the population, then a smaller number of solution in generation 20 are part of this overall Pareto front.
Thanks a lot 
