Multi-objective particle swarm and genetic algorithm for the optimization of the LANSCE linac operation
| Autor: | X. Pang, Lawrence Rybarcyk |
|---|---|
| Rok vydání: | 2014 |
| Předmět: |
Physics
Nuclear and High Energy Physics Mathematical optimization Rate of convergence Genetic algorithm MathematicsofComputing_NUMERICALANALYSIS Pareto principle Particle swarm optimization Multi-swarm optimization Instrumentation Linear particle accelerator TRACE (psycholinguistics) Domain (software engineering) |
| Zdroj: | Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment. 741:124-129 |
| ISSN: | 0168-9002 |
| Popis: | Particle swarm optimization (PSO) and genetic algorithm (GA) are both nature-inspired population based optimization methods. Compared to GA, whose long history can trace back to 1975, PSO is a relatively new heuristic search method first proposed in 1995. Due to its fast convergence rate in single objective optimization domain, the PSO method has been extended to optimize multi-objective problems. In this paper, we will introduce the PSO method and its multi-objective extension, the MOPSO, apply it along with the MOGA (mainly the NSGA-II) to simulations of the LANSCE linac and operational set point optimizations. Our tests show that both methods can provide very similar Pareto fronts but the MOPSO converges faster. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|
Full Text from ScienceDirect