[Freeciv-Dev] Re: (PR#10127) AI does not build either farmland or Supermarket

["Marko Lindqvist" <marko.lindqvist@xxxxxxxxxxx>]

<URL: http://rt.freeciv.org/Ticket/Display.html?id=10127 >

Jason Short wrote:
> A similar problem is that there's no incentive to build hoover dam until 
> sufficient factories are constructed.
> 
> The more exact our calculations are the more careful we have to be not 
> to fall into these dependency loops.

  Should we resolve dependency loops case by case or do we decide some 
hard rules? Like: In dependency loop between construction selection and 
infrastructure improvement selection, infra selection always takes the 
initiative? Hard rules (prioritized list of different selection 
categories, higher priority decision takes the initiative) would cause 
non-optimal solutions every now and then, but would make dependency loop 
avoidance easier.
  We have a lot of different decision categories and all of them 
potentially depend on all others.

  To list some main categories; some are not relevant for current AI:
(This list is in random order, this is not the prioritized list I 
mentioned earlier)

  - Tech selection
  - Diplomacy
  - Government selection
  - Production selection
  - Wonder construction
  - Infrastructure improvements
  - Military units
  - Worker/specialist handling
  - Tax rates
  - Intelligence
  - Exploration
  - Ferries
  - Victory strategy (space race/conquest/points/just try to survive)
  - Attack strategies
  - Defense strategies
  - Trade route creation
  - Pollution control
  - Capital placement
  - Counter-espionage
  - Settling
  - Happiness handling
  - Investments


  - Caz