More varieties of generations

Hello everyone,

I am optimizing energy consumption in a building using the Wallacei component. After analyzing the genome, I noticed that the optimization algorithm only considered a limited number of genes (or design variables). For example, I defined window-to-wall ratio ranges from 10% to 80%, but almost 90% of the genomes for all windows have a ratio of 10%. How can I define the algorithm parameters so that I can have a wider range of results? This means that I want more genes to be considered for the optimization parameters. I have attached one sample of the results, and as you can see, it forms an almost diagonal pattern. I wish to have generations with different results.


Wallacei does not just randomly shuffle through the variables to create a good result. It follows the principles of Genetic Algorithm (modified version for Architecture) thus it navigates the design space with intelligence. If you notice that the design variables are not fully explored, it might be due to the fact that those variables do not produce good and fit individuals so they are not explored more. How does Wallacei know this? By following the principles of evolution. However if you want to have more diverse exploration you can increase the mutation probability from 1/n to 0.5.

Yes, it makes sense. Thanks Milad for your response.

1 Like