Sounds like a testable release soon. GIMME TOYS!
Maybe, haha. The screenshots make it look like it's farther along than it is.
This economy simulation sounds really nice, even if it does have some interesting quirks.
It is still in barely-functioning early stages, but hopefully it will develop into something interesting.
I did some work with adding in some basic plots (intrigue-fueled plots, not plots of land) today, and having them trigger. I have one test case working in a very basic sense. I hit the same wall as when I tried working on basic government AI: Basic memory of things that we have already decided. For example, my test plot code looked something like "if we're second in line to the throne, plot to kill the first in line". That looks simple enough - they'd go and start a plot to do so. However, the next month, they would look at that condition, and start a plot to kill the first in line even though they already had started one previously.
There's a few options to deal with it that I can think of. There can be a binary switch called is_plotting which can be flipped on once they start the plot, and checked to make sure they only do this once. The second is to limit plots to 1 per character, and check to see if there is a currently active plot associated with that character every time the code rolls around. The first option means there can be tons of these switches to keep track of many things the AI does, which seems really messy to me. The second option won't apply to all cases.
Part of the issue is that this is going to extend to any kind of goal-driven behavior that the AI takes (raising an army, starting a plot, changing taxes) so I want to get this right. Does anyone have any suggestions, tutorials, or articles on goal-driven AI?
Author
Topic: Iron Testament - an ancient "open world roguelike" (pre-alpha) (Read 81986 times)