Shape Optimization of Multistage Depressed Collectors by Parallel Evolutionary Algorithm

摘要

In this paper a novel parallel meta-heuristic algorithm called MeTEO is presented, applied to the shape optimization of multistage depressed collectors, simulated by means of a Finite Element collector and electron gun simulator, COLLGUN, which uses the Constructive Solid Geometry for the description of the device shape. METEO is a hybrid algorithm composed by three different heuristics: FSO (Flock of Starlings Optimization), PSO (Particle Swarm Optimization), and BCA (Bacterial Chemotaxis Algorithm); it performs the optimization using both the topological and the metric rules and offers a natural parallel implementation that allows speeding up the whole process of optimization by the fitness modification (FM).