Timeline for What are the basics of implementing a Goal Oriented AI?
Current License: CC BY-SA 3.0
6 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Feb 5, 2017 at 22:15 | comment | added | Marc Jacob | I see, thank you for your help, it is invaluable. For now I think I understand everything, I'll ask more questions here whenever it's not the case anymore. | |
| Feb 5, 2017 at 20:53 | comment | added | Theraot | @MarcJacob That's the idea. To expand on that: Consider the set of action not fixed, since your pathfiding will have to see what actions can be done you may have a routine to discover the actions. Following the example of the Sims (look for articles about The Sims AI): your code may look to nearby objects, bookmark those that are relevant. Then query the bookmarket objects for actions, that give you your set of actions. About bookmarking locations, I'd suggest to play Globulation 2 (its free) it will make the idea clearer. | |
| Feb 5, 2017 at 9:10 | vote | accept | Marc Jacob | ||
| Feb 5, 2017 at 9:09 | comment | added | Marc Jacob | Read it, very interesting and helpful. The "Abstractions" part interested me the most, but not sure I understand it fully. So basically, for example, instead of trying to go for a plan like "MoveToCoordinates(whatever) -> DoActionOn(TargetObject) -> MoveToCoordinates(whatever) .." I should go for "MoveToTarget -> DoActionOnObject -> MoveToTarget" and let a subsystem deal with what is the target location every time and what is the target object ? | |
| Feb 5, 2017 at 9:00 | comment | added | Marc Jacob | Dang ! Thanks a bunch mate. I'll get to reading this wall of text you took the time to write for me. | |
| Feb 5, 2017 at 5:41 | history | answered | Theraot | CC BY-SA 3.0 |