The great post of Dinosaur Artificial Intelligence

Ok, so here's my review of the current state of the dinosaur AI as of 20/02/2019 :')

Common behaviors (displayed by all animals):

  • Resting. 50%
  • Socializing, which should include at least:
  • Play. 0%
  • Grooming. 50%
  • Dominance. 90%
  • Calling/Rejoinning. 0%
  • Panick, which should be “reseted” while in forest cover and express in either two ways, as in nature:
  • Fight. 50%
  • Flight. 25%
  • Hunger/Thirst. 90%
  • Stress. 25%

Carnivore behavior only:

  • Scavenging. 90%
  • Hunting. 50%
  • Pack-Hunting. 0%
  • Territorial behavior. 0%

Herbivore behavior only: I’ll describe mainly three, similar yet different behaviors I’d love to see: Herding, Flocking, Packing.

  • Herding/Packing. 50%
  • Flocking. 25%
  • Grazing. 0%
  • Stampeding. 0%
 
Fantastic post. I really, really hope these things are going to be added, since they would add so much more authenticity to the dinosaurs. Right now they're cool to look at, but after a while it gets boring. They should be able to do more things to make them feel more alive!
 
Your Feature Request / Idea
Hi folks,
Well, since it’s seems the most requested feature and, in my opinion, the most needed feature of the game now, I’ll try to summarize everything that was requested, plus adding my own points.

All behavior stems from a certain need or needs and, given the current dinosaur “needs” in game (which I think are good enough to implement some good, realistic behavior), we have some good ground to begin with; although, we still would need some kind of hierarchy within any given group of animals for certain patterns.

I think that, initially, this hierarchy would come pre-set as the order in which they came out of the hatchery (as in JPOG) but eventually, when reunited and having displayed the first interactions, said order could and should change to a more stable, logical one.
I think we can separate all behaviors for the sake of order so, first, some definitions:

Common behaviors (displayed by all animals):

Resting. I forgot this, but indeed is some of the most important behaviors that all dinosaurs should have. Not necessarily "sleeping", but just lying on the ground and rest. It would help a long way towards immersion for they currently look too mechanic, too robot-like, without ever getting tired. This would definitively require new animations.​
Socializing, which should include at least:​
Play; something most gregarious animals do: mostly chasing each other, bickering, chirping, biting. Could likely require new animations.​
Grooming; something some gregarious animals do, when able to establish some kind of bonding, de-parasyting each other mainly; some of the more intelligent could have some other affection displays as rubbing their muzzles. Would surely require new animations.​
Dominance; typical social behavior of most gregarious and non-gregarious animals, mainly the way to establish some hierarchy within the group. This wouldn’t necessarily be the non-lethal fights we have now. Fighting systems needs to be reworked, but that’s maybe for another post; here, let’s suffice it to say that a dominance display could as well consist on a “roaring contest”, and leave it at that. For gameplay, useful to establish an “alpha”.​
Calling/Rejoinning; whenever an individual gets too far away from the rest of its kind, it usually calls out for its pack/herd (if high enough in the hierarchy) or go running to rejoin them (if low in the hierarchy). The “calling” would also work among a proper herd, just to make oneself acknowledge by the rest; and this, in turn, could replace the gruesome chatting circles currently displayed, specially, by ornithomimids. This “calling” would be different in sound to the “alarm cry/call” refered to in “herding” and “stampeding” points (see below).​

Panick, which should be “reseted” while in forest cover and express in either two ways, as in nature:​
Fight; should the animal is big enough, naturally defended enough (armoured) or made more aggressive (here could have some influence the gene splicing) and the “threat” is not overwhelming (no human intervention, no bigger animal, no outnumbering).​
Flight; “default” and natural response for most herbivores and small animals, specially when outnumbered. Just g out and run for life.​

Hunger/Thirst: This should impel the animal to always tend to go towards the nearest source (feeder, water source) available, although not implying it should left the pack or abandon its own/entering another’s territory unless reaching a certain hunger/thirst level.​
Stress; typical behavior from caged animals on zoos. This should be apparent when the comfort levels drop from a certain point. It makes all animals look downhearted and nervous, wandering aimlessly, restlessly, crying a lot and growing weaker or more prone to diseases, not necessarily and immediately going to fence-attacking but could go for attacking other dinosaurs instead (in fact, some dinosaurs like ornithomimids should never attack fences, but could just run around aimlessly and alone). Since this would exhibit some “illness syntoms”, I think it’s already portrayed nicely enough currently.​

Carnivore behavior only:

Scavenging. Hunting is a most demanding behavior not all animals are willing to do. Tracking, stalking and chasing a prey is so hard that most carnivores, specially sick and old ones, would be happy to take a chance to eat off a carcass instead of go on any lottery-try to get fresh meat. Hungrier and weaker carnivores should always go for carcasses first, unless a proper feeder (which supposedly has fresher meat) is nearer and active.​
Hunting. Carnivores should, ideally, not try to openly go for any kind of prey, but try to approach them from the nearest forest area instead. Maybe this could be done with some map coordinates, as the maps are always the same. Otherwise, if hungry/interested enough to try and openly go for a prey, it should dynamically change target to the nearest prey available once their flight begins (although I admit I don’t know how they select a prey currently). Also, they should not have a 100% rate of success: healthier preys could prove fastest, and they are running for their lives. Key point is, whenever a hunt is made, the death dinosaur keeps “linked” to its slayer until totally decaying, thus impelling it to feed from it, stay near and protect it from other oportunistic carnivores/scavengers. It would also be nice if the bigger carnivores could drag smaller preys to within the limits of their territories too (see below).​
Pack-Hunting. This is one of the key points. I think it should start somewhat similarly to the stated above, but with individual needs calculated in a “social” way: the higher the individual is on the hierarchy, the most relative importance would have it’s hunger need; so, when the alpha wants to eat, it would call for the pack to move; if the beta is hungry, but the rest of the pack is not, perhaps it would need for any of them to reach a certain hunger level, for the alpha to call the pack-hunting, and so on. That would also mean that, if the lowest on the hierarchy reaches a certain critical-minimum hunger level, alpha should call immediately for pack-hunting, as it would reach the “social” hunger level, even if the others aren’t hungry. Not sure if I explained myself, but will gladly elaborate. Only exception: should the weight difference between individual predator and prey is not that high, then individual hunt could trigger there for an individual hunting need. Yet, that could (and should) also trigger dominance displays and territorial behavior (see below) over the carcass, during the aftermath.​
Pack-Hunting dynamics would highly feed on social limits (thus a “pack” already “complete” wouldn’t accept any new member) but would also share territorial behavior (see below), so that, given enough space, two established packs could also coexists and clash on occasion (although I think that, lore-wise, Raptors alone should not tolerate this).​
A note about the "Thirst" need: this would be the only one totally "individual", although it would be limited to the "herding/flocking" dynamics (see below) properly applied to such a hunting-pack and the only one to make the whole group move, regardless of who calls for it, aside from the "panic" state.​
Also a note on feeding, as pointed out by AcroCarnoRex: the hierarchy of the pack would also determine a "pecking order" so that, when hungry, the "alpha" eats first, whatever the food source (carcarss, feeder, prey), and only when he is full, allows the "beta" to eat. Huge discomfort levels from lesser members of the pack could and should lead to dominance fights which could or could not win, depending of their health, hunger level, aggression, etc​

Territorial behavior. This would mean two things, somewhat intertwined with the above: first, according to the animal’s space needs, and once it has roamed/(scanned) the whole paddock area (and established a pack, if it can), it should define some small portion of it (that still, more accurately, meets its needs, particularly the “water source” things) as it’s “territory”, thus keeping mostly within its limits. I am not sure it this is doable, but I think it could be, by attaching some “unseen” coordinates of the island to the animal: territories became usually inherent and unchangeable, and most animals would stay within their known limits for safety, unless severely drawn out (usually by hunger, easy hunt, etc) True, “territories” wouldn’t be regular circles or geometrical figures, thus making said coordinates difficult to establish, but I don’t think it needs to really be that strict: just a “space within space” big enough to be comfortable in and mostly staying within; it could even “contain” some area “outside” the fence, thus making it more likely that, should the fence suffers any damage, the animal escapes towards it.​
This could easily mean that, if the paddock is twice the size needed for a Tyrannosaurus, it could actually house two of them, with a random chance of both clashing while establishing territories (thus displaying Dominance behavior/fighting) and/or, having the loser stressed should his part does not meet its needs (for instance, if his area doesn’t have enough forest or has no feeder/water source within it) or if it just happens to be more aggressive than the first. Any attempt from other dinosaurs to entering it, should lead to dominance displays, chases and fights depending on the attack/defense stats of said dinosaurs.​
Second, this would also mean some kind of “marking behavior” (peeing, brushing oneself over the ground or against trees, roaring…) to keep stating/claiming that territory’s property. That would need, of course, some new animations. And I think it even plausible that herbivores and smaller animals keep themselves well away from the bigger carnivores territories too thus, when not properly fed, big carnivores should venture out of their limits to hunt.​

Herbivore behavior only: I’ll describe mainly three, similar yet different behaviors I’d love to see: Herding, Flocking, Packing.

Herding/Packing: I link them both because they would work in the most similar way, making all animals stay close to their kind. I think there needs to be a minimum distance between animals of the same species (even as an “unseen” stat) aside from the social/population limits. That distance would vary according to the size and social needs of each species and would slightly differ between “herding” and “packing”: while it would tend to keep animals of the same species close in both cases, “herding” animals would tolerate animals from other species among them, while “packing” species would not, thus making them more prone to gather themselves and separate from other species. That would effectively show those lonely families of Ankylosaurs, Trikes and other unconcerned armored herbivores wandering somewhat separated from the true herds of Hadrosaurs.​
Flocking. This, I think, would be a particular instance of “herding” which would differ in that all animals should move together (in the same direction), specially while “panicked”. Thus, while resting, socializing, eating, etc. the flock would behave quite like a tight herd, but turning to a “pack” while in panic state. Which leads us to the last one.​
Grazing. (Thanks to AcroCarnoRex for pointing this out) Grazing is a huge part of the daily doings of many herbivores and, specially the biggest ones, would need huge amounts of plant material just to keep going alive. So, the fact that many Sauropods, for instance, can afford to wander far away from trees and feeders is a bit odd. What I think should the done to improve the current system: herbivores should be eating almost constantly, while not playing, resting or grooming, thus all veggetables within their enclosure should be susceptible of becoming "food". Now, there's also the JP/JW canon of the Lysine dependency so, and even while this (almost) constant grazing would indeed decrease their hunger, it wouldn't even be enough for it to keep from lowering so, eventually, they would starve without the proper assistance of their respective feeder. That would also function as a nice way to make the feeders' timing longer, as they wouldn't need them so often. Also, should some kind of new flora is implemented at some point, it can be "marked" as "favorite" by certain species, thus luring them towards the patches of it; this areas would have some inner cool down, that would render them useless after the herbivores have eaten over them, making them "free" to move to another patch or try find a feeder if they are still hungry, all the while faking some nice "migratory behavior".​
Stampeding. This should be something all social herbivores should do, specially Hadrosaurs: whenever one of them feels “threatened” (be it by spotting a carnivore, the noise of an helicopter or some other stimuli), it should cry a “warning call”, more or less immediately (fastest in “flocks”, slowest in “herds” and “packs”) moving the group into action: fleeing. Now, “herds” would scatter away from the danger “source”, preferably toward forest cover all the while trampling and destroying everything in their course (be it feeders, lighter unelectrifyed fences or smaller animals), while flocks would just keep running away, changing direction almost at the same time, and “packs” would tighten themselves into a defensive group stance should forest cover lies too far away.​
Excellent idea!!!!
 
Fantastic post. I really, really hope these things are going to be added, since they would add so much more authenticity to the dinosaurs. Right now they're cool to look at, but after a while it gets boring. They should be able to do more things to make them feel more alive!
My point exactly. Although new behaviours would also add a lot of depth to the game-play too. Stampeding, pack-hunting, the calling/rejoining of herds and territorial behavior particularly.
 
Maybe one of the the most important Dino IA, is about the packing huntig just for the little carnivores, if we have the basics code,
it could be easily implemented. Also, it should be implement that small carnivores could climb the fences.
 
Last edited:
After the excitement of the new fish feeders and pachy self defense, I think we need some more social interactions to help make these dinosaurs quite distinctive from each other.

More in the sense that the new pachy self defense is only ever brought up during fighting moments, and a lot of the time that we would play the game, the dinosaurs would (usually) not always be fighting. So some more idle animations such as grooming between herbivores, more exaggerated sentry mode for hadrosaurs and sauropods, playful fights between small carnivores and pachys, and lounging for large carnivores. These new idle animations would add more life to the dinosaurs in a much more observable way.
 
Top Bottom