Hi everyone! New Stardew player here, but long-time optimisation geek. I've just gotten to the Greenhouse part of the game, and it seriously triggered that geekery.

I've seen a few discussions around optimising the profitability of the Greenhouse, but a lot of the analysis people have done focuses only on a single constraint - usually the space available in the Greenhouse. In reality, we are likely to have a lot more constraints than this in play - how many kegs, jars, and casks we have, and how much time we want to spend, and some of the existing answers, particularly promoting blueberries, simply assume all these other resources are unlimited.

Having encountered Operations Research and formal optimisation methods, it occurred to me that this is a classic "product mix" problem with multiple constraints. I decided to write up a simulation of this in MiniZinc, and have it tell me the optimal mix of crops in any given situation.

The results are as follows:

There are a few somewhat game-breaking crops that are acquired via non-conventional means. If you have Ancient Fruit, Starfruit, or Pineapples, you should make wine with them in that order, and age the most valuable wine you have. Once your ability to make wine is saturated, you should grow Sweet Gem Berries if you have them, to sell raw.

Ignoring those crops, and going only for more "normal" options, we have some interesting rules of thumb:

• When you have plenty of kegs, with labour in moderate supply, make melon and strawberry wine. The melons are somewhat more profitable if aged.
• When you have plenty of jars but no remaining kegs, with labour in moderate supply, make red cabbage and pickle it.
• In both the above cases, strawberries and blueberries can fill out your field to be sold raw, once your ability to make artisan products (i.e. jars, kegs, labour) has run out. Blueberries are more profitable to sell raw with Quality Fertiliser and lower, strawberries are better with Deluxe Fertiliser.
• With unlimited labour and unlimited jars or kegs, blueberries are effective in both jars or kegs. However, blueberry wine is significantly less effective than melon or strawberry for ageing in casks. Note that even without ageing, we should switch blueberries for strawberries if you value your time at more than 30 coins per "action" (planting, harvesting, loading/unloading jars/kegs/casks). In turn we switch the strawberries for melons at about 100 coins per action.

If anyone wants to play with the model, or find where I went wrong, I've put them in pastebins:

• All of these are written for the "MiniZinc" optimisation modelling tool.
• My MiniZinc model file is https://pastebin.com/EfzEwh7S .
• My MiniZinc data file for an example situation is https://pastebin.com/uA75BFnQ .
• Select the "COIN-BC" backend engine in MiniZinc, as many of the others are seriously broken.
• All data comes from https://thorinair.github.io/Stardew-Profits/ and was extracted using statements like those in https://pastebin.com/32XHirsn .

To run your own situation, look at the settings at the end of the "data" file.

I hope this is useful to some optimisation nerds like myself :) Corrections, comments, etc all welcome.