Boids | Flocking behavior simulation

After a few days of struggle I finally got it working. I got the idea from this paper. Particularly this sentence:

Each boid considers its distance to other flock mates in its neighborhood and applies a repulsive force in the opposite direction, scaled by the inverse of the distance.

The code bellow does exactly what the above sentence says.

You really should not include those folders because they are dependent on your system and will most likely not work for other people (e.g., Mac vs Windows). Yes, a .gitignore (Note the spelling) is the way to go. There are many examples out there. Often when you create a Git Repo you can select to include one. However, it sounds like you already have a full git configured so it's best to push to an empty GitHub repo. Here's one that I often start with for Python projects: https://github.com/github/gitignore/blob/master/Python.gitignore

std::unique_ptr can't be copied, it doesn't have copy-constructor but has move constructor. You can use std::move to convert boid to rvalue then the move constructor could be used.

You can use std::find to check if an item already exists in a container. Or you could use a container that contains unique keys e.g. std::unordered_set.

You need to be very careful when storing addresses. If the object moves or goes out of scope the address becomes invalid. This is in fact what can happen in your example because std::vector will move objects around on resize.

Solutions:

• associate and store some unique identifier instead
• use std::unique_ptr (std::vector> boids;), this will ensure the object doesn't move (it's the smart pointer that moves)
• make boids vector or set const (initialize it on construction) and make sure the containing object (if any) doesn't move throughout your access via pointers.
• use a container that doesn't invalidate iterators on resize e.g. std::list

I tried doing it the std::unique_ptr way. It does not work when I try to push back the memory address

withinSensoryRange should be a vector of raw pointers:

The reason why you are not seeing obvious performance gains from the Quadtree is because of the nature of your simulation. Currently, the default separation causes a large number of boids to "converge" to the same position.

Lots of objects in the same position is going to negate possible speedups due to spatial partitioning. If all the objects are in the same or near position, boids in the area a forced to check against all other boids in the area.

You can demonstrate to yourself that your Quadtree is working by watching or profiling your application with its default settings. Now turn separation up to maximum. You will see visually, or through profiling, that as the boids spread out more evenly, the FPS increases significantly. This is because the Quadtree can now prevent computations thanks to its spatial partitioning.

With default low separation:

With maximum separation:

You can see how performance is increased in the second image. Also note, that the conjecture by another commentor that it is the construction of the Quadtree (insert) that is taking up all the time is wrong.

While in some applications you may be able to update a Quadtree as things move around, since in this simulation every constituent moves every frame, reconstructing the Quadtree from scratch is less work, then taking every object out and reinserting it into its new position.

The advice to skip square-rooting and just use the distance-squared is good though as this will get you a bit more performance.

That does not look like a proper rotation to me. A quick search for a 2D rotation matrix yields something more like this:

From your demo, the z component is flipped, and you can test this from only trying one of the rotations at a time. Second, chaining rotations in 3D this way will usually not do what you want, as rotating will change the "up" or "right" vector of the coordinate system attached to a certain object. For example, rotating about the up (-y for p5) vector, or the yaw angle, will rotate the right vector. The second rotation then needs to be about the rotated right vector (now pitch), so you can't just use rotateX/Y/Z as they are still in world space instead of object space. Note that I'm completely ignoring roll in this solution, but if you look at the boids from the front and top angles, it should be aligned with the velocities

If each entity needs to know about each entity, then it seems like it's not the entity that's responsible for the update. You should probably refactor from a method on the entity to a function that operates on all entities.

Iterating over the entities in a nested loop is a bit difficult. One solution would be to iterate over one copy of the array and allow the other loop to take a mutable reference.

Note that there is a little more detail in a 1999 GDC paper. There are also many open source implementations, such as this (in retrospect, surprisingly complicated) one in OpenSteer.

An agent’s seek behavior simply assumes a desired velocity toward a static target, and defines a “steering force” which is the error/difference between that and its current velocity.

So a pursue behavior requires only to estimate a current target location. A rough approximation is usually good enough, since it is discarded after this animation frame / simulation step. Typically steering behaviors use constant velocity predictions, simply multiplying the current velocity by some time interval. A helpful estimate for that interval is the distance to the quarry agent, divided by our current speed.

How this interacts with (a) the pursuers being a flock, and (b) a simulation based on path finding, would depend on specific details of your game. Probably you want to blend a small pursuit “force” with the flocking behavior so it does not interfere with the nature of the flock.