Add boundary handlers for stuck particles

Is your feature request related to a problem? Please describe. There are some stuck particles as reported in #234 and #150 . This is related to compute_position waiving position updates whenever a particle will potentially end up out-of-bounds. There are standard ways to avoid this problem. The reflective bounds is one option, but the related PR is currently stale, so it may be important to create a general framework for handling the boundary problem.

Describe the solution you’d like Create a set of handlers, such as a BoundaryHandler. Usage example:

Usage outside of the class

# Usage outside of the class
from pyswarms.backend import operators as ops
from pyswarms.backend.handlers import BoundaryHandler

bh = BoundaryHandler(strategy="reflective") # And other arguments here if it's parameterized
ops.compute_position(swarm, bounds, handler=bh) 

Inside BoundaryHandler

# Inside BoundaryHandler
class BoundaryHandler(object):
    def __call__(self, position, bounds, *args, **kwargs): # parameterized
        # Some logic here depending on the type of handler, for example:
        # (might be better as a dictionary with a string as key and callable as
        # its value)
        if self.strategy == "reflective":
            new_positions = self.reflective(position, bounds) # parameterized

    def reflective(self, position, bounds):
        """Some docstring here"""
        pass

How it will work in the backend:

if handler:
    new_position = handler(swarm.positions, bounds)

[whzup]

Yes maybe we can work on the same PR branch?

Yes! I have never done that, might be a nice experience 😄

There’s also the need to use handlers for Velocity

I wonder if it might suffice to use a VelocityHandler. If the particles don’t even leave the bounds there is no reason to fix the position 🤔 Maybe we can try it out like that?

[whzup]

Ok 👍 Let’s do this one!