Proliferator Mixes

[Martin-Pitt]

It turns out proliferator mixes may be a lot more complex.

That I can't apply proliferator presets per recipe necessarily — some setups may actually have solutions for different setups for the recipe's chain.

So, bruteforcing for the entire recipe chain basically.

Additionally I would like to generate proliferator mixes by code, this is so that new recipes can be handled automatically and also because most mixes aren't complete, particularly they do not cover buildings at all.

There are also additional mixes I would like to create, particularly:

• Optimise Facilities
  • Effectively optimise the total amount of assemblers, smelters, etc.
• Optimise Facilties + Sorters
  • We can make an assumption that each item and result would require one sorter each to add to the total buildings count. A few recipes are actually affected by this.
• Optimise Facility Footprint
  • We can assume that some facilities occupy a certain footprint, minimally at least the size of the building itself on the grid. Some proliferator mixes may be more optimally leaning towards smaller smelters compared to the large assemblers

[Martin-Pitt]

• Code to bruteforce proliferator mixes has been implemented. It can optimise statistically for facility & sorter counts.

• Pregenerated mixes would be one way to go. However it does get complex with recipe preferences adding more layers to the possibilities which are already very high. The most complex recipe, Universe Matrix, is already impossible to calculate in a reasonable time. Also can preferred buildings affect the total (facilities + sorters) optimisations enough to sway mixes?

• Need to redo the calc output rendering to use the new architecture.