Skip to Main Content
This paper introduces a novel particle swarm optimization algorithm based on the concept of black holes in physics, called random black hole particle swarm optimization (RBH-PSO) for the first time. In each dimension of a particle, we randomly generate a black hole located nearest to the best particle of the swarm in current generation and then randomly pull particles of the swarm into the black hole with a probability p. By this mechanism of random black hole, we can give all the particles another interesting direction to converge as well as another chance to fly out of local minima when a premature convergence happens. Several experiments on fifteen benchmark test functions are conducted to demonstrate that the proposed RBH-PSO algorithm is able to speedup the evolution process distinctly and improve the performance of global optimizer greatly. Finally, an actual application of the proposed algorithm to spam detection is conducted then compared to other three current methods.