Decisions, need advice

[NiclasCage]

I'm wondering if the following is a system that
a) is possible to create in the first place
b) the computer will be able to use

I want nations to be able to increase/decrease a number of their tech-teams' skill. This should cost money (or refund some of it).
I don't want tech to unlock new models anymore. I want tech to unlock multi-choice decisions, where you decide how much money you want to "throw" at the new model.
After a certain amount of time you unlock a new model based on the latest technology. The decision is then put to sleep until the next technology level is researched.

I want some randomization here, both in terms of days taken to develop a model, and the quality of the model. I suppose I have to design several models for each tech level.
While I don't want hopeless or very overpowered models, I do want a system where you may think "Oh, this tank is good", or "Damn, this submarine leaves a lot to be desired".

I don't know what limits the game puts on me here, both in regards to modifying skill levels by decisions, the total number of models, and making the AI use it to a certain degree of non-stupidity. Any answers will be appreciated.

 

[Altruist]

I want tech to unlock multi-choice decisions

b) the computer will be able to use

Well, as soon as "choice" is included...
# in the sense that one choice is better than another dependent on various variables... you must be either able to script those dependencies in a way that the computer can make a good choice or the computer cannot use it
# the exception would be if there were choices but it doesn't really matter what is chosen (which usually brings up the question what kind of choice this is... but there might a place for RP, flavour, distinction)

Misc:
Instead of trying to "teach" choice to scripts, there is a workaround: mimick choice.

tech to unlock multi-choice decisions

Have a certain number of endmodels which could be reached by going that multi-choice tree to the very end.
Throw the dice to find what submarine-type, what tank-type this empire should strive for in this game.
After this is decided, every tech decision is no longer a decision but a path which needs to be taken to get to that subarine-type, that tank-type.
This forces you also to do a very important thing: defining your multi-choice-decisions.

Later, much later, you might even add something like:

country doesn't own a single drop of oil => go infantry way
or
lots of oil => go tank way
else random (landunit-model)

owns more than 10 battleships AND Fleet-in-Being => go for crazy BB-design
or
owns zero BBs AND Indirect Approach => go for sub-desgin
else random(ship-model)
...