Skip to content

Latest commit

 

History

History
26 lines (17 loc) · 2.47 KB

README.md

File metadata and controls

26 lines (17 loc) · 2.47 KB
Raw

SatisfactoryLP

Satisfactory linear programming optimization, using scipy.optimize.milp.

Recipes are parsed from data provided by Coffee Stain Studios in the Satisfactory game directory.

Map data for resource node availability is taken from Satisfactory Calculator. I don't know how to obtain it directly.

Data sources, retrieved 2024-11-04:

  • Docs.json: Satisfactory\CommunityResources\Docs\en-US.json (v1.0.0.5 - Build 377620)
  • MapInfo.json: https://static.satisfactory-calculator.com/data/json/mapData/en-Stable.json?v=1730198931 (plus auto-formatting)

Baseline results on Google Sheets, generated 2024-11-04.

Requirements

  • Python 3.12 (older versions may work)
  • pip install scipy xlsxwriter

Running

Running python SatisfactoryLP.py produces some text output and a Results.xlsx file which can be opened in Excel or imported to Google Sheets. There are various options to control extra costs and penalty weights; use -h to see details.

The most important options are:

  • --machine-penalty controls the tradeoff between underclocking and machine count. Increasing this penalty causes the optimizer to prefer higher clock speeds and fewer machines. On the other hand, decreasing it to zero would cause most machines to be set to the lowest possible clock speed.
  • --manufacturer-clocks specifies the allowed clock speeds for non-somerslooped manufacturers and Water Extractors. Increasing the granularity allows the optimizer to reach higher objectives, at the cost of increased problem size and running time. All clock speed options can be specified as a comma-separated list, where each token is either a single clock speed or a triple lower-upper/step. The default setting is 0-2.5/0.05 which means "every clock speed between 0% and 250% at 5% increments." This is somewhat overkill and results in 51 versions of each recipe being created; however, it still runs in under 3 seconds on my machine. For faster runs, a reasonable setting is 0-1/0.25,1.5-2.5/0.5 which allows clock speeds [0%, 25%, 50%, 75%, 100%, 150%, 200%, 250%]. This runs in about 0.7 seconds and the resulting objective is only 0.12% worse. (Note that 0% will automatically be forced up to the minumum clock speed of 1%, and higher clock speeds will automatically be forced down if they would exceed a conveyor belt limit for the recipe.)

TODO: document more options.