|
To Firaxis: Use simple Marginal Utility formula to Improve AI effectiveness
I'm not sure if it is possible to include this in a patch as I'm unsure of how the AI is built.
But I think there is a very simple and effective economic idea, marginal ultility, that can make the AI more efficient in deciding what to do capture, and what to defend.
The idea of marginal ultility is well explained in any introductory economics text, but it essentially says that people attach values to things. If we have a lot of one thing, getting an additional unit is worth less to us than getting an additional unit of something we have little of.
Applied to the game, this concept can be applied in two ways. One is city management (how much of one unit should be built?) or if that is a bit too complicated, apply it simply as a function of the AI's strategic decision on the overall strategic landscape. This can be applied so that important cities for the AI will be held more vigoruously, important lone resources for the AI will also be held in similar fashion. And the AI knows which it has a lot of and which it has a few of and make appropriate decisions as to what to obtain, or what to deny you. If it spies on your civ and finds out you have one source of iron, it may make it an objective to take that source or disrupt that supply.
The beauty of the marginal utility idea is that is can be expressed as a mathematical equation, and the co-efficient is a simple and managable number. The AI can decided what to do by simply running a routine to compare the relative size of these numbers and deciding from there which action is the most important.
|
