Plan
Principles of the "simplest" PSO
Moving the swarm (no-hope/re-hope)
Results with two "toy" examples
Notes:
The first idea is to define an hybrid method as a combination of PSO and one or two others, so that we can take advantage of several approaches.
The second idea is to begin with a very simple algorithm, just in order to see how good or bad it is, and to improve it step by step. There is here an underlying notion of cost: a complex algorithm may be better, more efficient, than a simple one, but it
Then we will see, in this particular case, that the swarm can be moved in two ways: the "normal" move and the "re-hope" move, when it appears there is probably no hope to reach a solution.
Two toy examples will be shown. Results are pretty good, but, of course, it doesn't mean such a method would be very useful for real problems. We will discuss this point, in order to define what could or must be improved.