Parallel particle swarm optimization can solve many optimization problems quickly on GPUS
Erik Wynters · Journal of computing sciences in colleges · 2018
Particle swarm optimization simulates the way a flock or swarm of birds, fish, or other animals move through space and uses that to search for a good solution to a problem. This optimization method is easily adapted to a variety of problems and is easily parallelized since the same calculations are performed for each particle in the swarm. For many problems, a parallel implementation run on a powerful graphics card's multi-core processor (GPU) runs hundreds or thousands of times faster than a serial implementation run on a CPU. This is shown by using parallel particle swarm optimization on GPUs to solve several different facility location problems.