GidMod, Status Update... |
Well, so far so good. Extensive testing has uncovered some more errors in my mod which have been corrected (such as the AI being able to spend over 100% of its intel points on sabotage actions under certain conditions).
Currently, however, there is still much work that can be done. As the mod continues from this point forward, one of my ongoing tasks will be to build ways to further differentiate each race's behaviours from each other.
I would say that, at this point, Research is the weakest part of this AI. I can not add any more tech groups, as I've hit the limit the UI code will allow. If I add any more, errors occur in the Research window. Therefore, to get past this roadblock, I need to completely rethink how the AI goes about research, and implement a different system than what exists now. While this will (hopefully) give the AI more flexability in its research, as well as making it more dangerous, this will be a time consuming task.
I'm shooting for a release date of this version on Friday/Saturday, so the Research changes will have to wait.
In the meantime, the AI does seem more aggressive than before, but I will need feedback from the players as to whether it is too much or not enough. Its all done with variables and numbers now (rather than categories and treaty status) so this can be tweaked endlessly (code permitting). Players should note that the AI is more aggressive towards those it is actually at war with, so treaties still matter. However, the AI will also perform aggressive actions against those it is NOT at war with (even those it has treaties with) if its anger is high enough. So, watch that anger level as an indication of what actions the AI will take. In a later version, I'll include ways the AI empires may try to "butter each other up", in order for certain personalities to try to create coalitions (and include code so that AI will form full-on alliances without player intervention). Then comes the "ganging up" scripts, which will cause the AI to prioritize its enemies, check for others relations, and try to form coalitions against those it feels are its greatest threats (even amung races that are otherwise enemies). All of this is do-able, it just takes time.
The next version will probably focus on completing the changes to AI ship design, so that each race will design ships with different priorities in mind. If this goes quickly enough, then I should be able to include changes to how the AI builds combat ships, so that it will try to have a mix of ship sizes. This mix ratio will be designed so that I can set it by each race as well.
I'm going to try to push through the changes to AI ship design and construction as quickly as possible, so that I can get on to redesigning the research system ASAP. After the research system, it will be time to augment/adjust the personality/politics system.

Re: GidMod, Status Update...
Either way, since I won't start writing my mod until 2 years later I look forward to what AI inspirations you and the other modders come up with.
---Sig---
Playing Touhou games (Go to this Wiki if interested. No, nobody else is that good/insane as that replay). No rush for SE




Re: GidMod, Status Update...
Incidentally what are the challenges/limitations of coding AI Research? That was a huge part of what killed IRM's code writers to my understanding.
If I were to write a tech tree from absolute scratch, what guidelines should I follow to make it "AI friendly"?
---Sig---
Playing Touhou games (Go to this Wiki if interested. No, nobody else is that good/insane as that replay). No rush for SE