IslandPlanner

[dragonheart107]

didn't want to spam PRs for now, so created this

322d06d

Long config:
I found it difficult to have a clear choice for selection and delegated the decision to the users:
A character may have a high attribute in one attribute and based on that would be able to work in multiple different places, but due to skill, should be used in another specific place, e.g Belfast/FdG should always work in Cafe for cheese etc.
also level of skill may matter, for example laffey's bonus, will make her the best choice for most at max level
also since ranch item scarcity has quite a gap (e.g eggs vs milk), i decided to split it into 4 instead of a single pool

Using chars in production is mainly for those with multiple chars.

As for transport, it was mentioned that it wasn't implemented due to uncertainty on how to weight offers:
Is something like
exchange for specific items (e.g Milk, poultry, crab/squid (for paella)) >>> most pt profit >>> most coin profit
not good enough?

[guoh064]

Some comments on your last question:

The original idea of island production module is to make sure that value of items are calculated from their natural needs instead of human-defined prices. This should be robust in the meaning that if one day milk becomes very cheap (either due to larger batch size in the same time or due to fewer needs for milk products like cheese and cheese products), we do not need to rebuild the configs, but just need to rerun the calculator against the new data. A human-determined config harms the integrity of this design.
Such needs are derived from linear programming and are implemented into daily buffer numbers, which avoids the necessity to make a schedule for production. For production the only standard (if simplified) is the level of storage against daily buffer number.
The weight of transport orders, under this idea and under current config, is to evaluate how it fluctuates the storage, and in particular change the amount of items, per the daily buffer. For consuming part, it is almost always safe to deplete redundant items and dangerous to use those in shortage. For reward part, it is welcome to replenish items in shortage, but you cannot say an additional batch of redundant item is not useful (it may contribute to island points). So the evaluation function is not the same (or the counterpart) for both parts, which means the evaluation should be taken care of.
The aforementioned strategy also relies on the availability of knowing the precise storage numbers for the reward part, which is bothersome at the moment since a scan of storage is required, otherwise you only get partial information and is limited to take the greedy solution.