PSO with Dynamic Topology and Random Keys method applied to Nuclear Reactor Reload

Abstract

This paper presents a new way of communication between particles in Particle Swarm Optimization Algorithm, called Dynamic Topology. In this approach, probabilistic methods are used to choose which particles will communicate and update on every iteration loop. Furthermore, a new idea of Global Best is created seeking optimize the process indirectly. Here, PSO with Dynamic Topology (DTPSO) and Random keys method was applied on a real problem of nuclear engineering, the Nuclear Reactor Reload Optimization Problem (NRROP). The results of the implementation were compared with others metaheuristic presented in the literature and have shown that the methodology proposed acquired better solutions.