Subject
so, guys..

Text
: i think you're both right and good way of mixing these two ideas might be: : : say the 'overall bigger picture' goal is a whole lotta research. the "government" passes down this order, and each city gets it. then, specified somewhere, is what a city can do in order to help achieve this overall goal. it might include several options such as building certain buildings, manipulating tile exploitation, etc. so the city can then decide which one of these it wants to do, based on which will have the most benefit for this city, AND the least impact on other things that the city is already doing and managing itself, ie military needs or city growth. : : the issue then becomes 'balancing' out all the issues. basically you would need some kind of rating system where the 'overall goal' is specified by, say, "an increase in overall science by x amount". then you could calculate a way to do it that has the least impact on reducing other things such as overall production from all your cities, overall growth, tax rates, etc. : : other things that arent as simple as 'reduced food by 3' might be things like cancelling a building its already building in order to build something else to reach a newly set 'overall goal'. there you would have to take into account how much of the building was already built, how long the whole thing would take, and what benefit it would have should it be completed, in order to assess the negative impact of stopping it midway in order to start something else, and compare that loss to the gain of the new project, time to build being taken into account in both cases. : : so, all in all, it would calculate ways of reaching the goal, each with their own level of 'negative impact' on other things, and then it could simply apply them in order of low impact ones to high impact ones, one at a time until the 'overall goal (increased science by x amount)' had been satisfied. : : thats kinda what we were getting at then, right? : : -- : : the other question put forward was, what method do current successful AI's use. well i can tell you based on my experience, the previously highly successful military-based AI Civ-Seed used the method of either being in 'normal' mode, where each city was basically trying to grow in population, and buildings being built just depended on what it didnt have yet, with an order of what to build first.. and then 'military' mode, where every city put EVERYTHING into production tiles, even if it meant the city was LOSING food (although they wouldnt let the population go down), and all production was put into military units. this meant they could produce their army at an insane rate, which made them a dangerous foe. : : another example i can think of is KIAI, which basically, for each city tries to get the most of everything, not considering individual rates of things, and builds all its buildings in a pre-set order, and then once it has nothing left to do it switches to either military or workers, based on i dont know what, might even be random, they never seem to stop doing either. thats how it seems to work anyway. : : -- : : hope that helps?

Administration

Password