In this article a new simple-structure variation of Particle Swarm Optimization (PSO) algorithm is proposed. Since the standard PSO has a very good performance, the new variation retains many properties of standard PSO such as stochasticity and some other properties similar to those in Evolutionary Algorithms. However, unlike the standard PSO algorithm, in the new algorithm the particles can not only communicate with each other via the objective function but, via a new variable named “quantizer”, and hence, the new algorithm is labeled as the Quantized Particle Swarm Optimization algorithm. In addition, extensive simulations are given to show the advantages of the new algorithm over the standard PSO. Finally, the convergence analysis for deterministic version of the new algorithm is also presented.
Convergence analysis of evolutionary algorithms in the presence of crash-faults and cheaters