Comparison between Deterministic and Stochastic formulations of Particle Swarm Optimization, for Multidisciplinary Design Optimization
Daniële Peri, Matteo Diez, Giovanni Fasano · 12th AIAA Aviation Technology, Integration, and Operations (ATIO) Conference and 14th AIAA/ISSMO Multidisciplinary Analysis and Optimization Conference · 2012
Particle Swarm Optimization (PSO) is having a growing space in the optimization community, mainly due to its appreciable qualities of fast initial progress, reduced computational cost and parallel structure, suitable for High Performance Computation (HPC) platforms. Original formulation includes some random coe cients, so that a statistical analysis of the solution is often needed. To avoid the latter situation, Deterministic Particle Swarm Optimization (DPSO) has been introduced: removing all the random coe cients the DPSO is a deterministic algorithm, so that a single run is considered to evaluate the success of the algorithm. In this paper, a comparative study between PSO and DPSO is reported, in order to investigate the performance of DPSO versus PSO.