Particle Swarm Optimization Algorithm

Particle swarm optimization searches for a good solution to an optimization problem by maintaining a population, or swarm, of candidate solutions called particles, each moving through the search space with a velocity adjusted at every iteration to be pulled partly toward the best position that particle has personally found and partly toward the best position found by the whole swarm, so the swarm gradually converges on strong regions of the search space. It requires no gradient information about the function being optimized, making it useful for problems where the objective function is not differentiable or not known in closed form. James Kennedy and Russell Eberhart introduced the technique in 1995, inspired by the collective movement of bird flocks and fish schools.