基于遗传算法的自由电子激光优化设计  

Application of genetic algorithm for optimization design of free electron laser

作　　者：张白鑫 张彤[1] 陈建辉[1] 刘波[1] 王东[1] 

机构地区：[1]中国科学院上海应用物理研究所嘉定园区,上海201800 [2]中国科学院大学,北京100049

出　　处：《核技术》2016年第2期1-6,共6页Nuclear Techniques

基　　金：国家自然科学基金(No.11175241)资助

摘　　要：自由电子激光(Free-Electron Laser,FEL)的辐射功率、光谱等关键量是表征FEL品质的重要因素,这些量往往依赖于多种参量,所以优化这些品质参量的问题即可等效为如何寻求合适的参数来获得更优的FEL的输出。遗传算法是解决这类多变量优化问题常用的算法之一。本文基于遗传算法设计了一个用于FEL优化的应用程序,该应用程序利用实数编码方式,选择合适的算子并作相应的改进,同时利用Java Swing构建了友好的用户界面。实验结果表明,在进行辐射功率优化时,该算法能够在较短的时间内寻找到非常接近全局最优解的较优解。该应用程序具有良好的通用性与可扩展性,在一定程度上为FEL装置的运行优化提供帮助。Background： Optimization of free-electron laser （FEL） facilities is of great significance to achieve radiations of high-quality, e.g. higher brilliance, purer spectrum. In addition, these properties are affected by various parameters, e.g. the electron trajectory along the accelerator and undulator, the beam envelope or beta function, which could be changed by tuning the correctors and quadrupoles. Purpose： This study aims to find a set of suitable parameters to optimize the radiation power. Methods： The genetic algorithm （GA） is applied to investigation of the FEL power optimization with respect to the focus-defocus-focus （FODO） lattice configuration between the undulator segmentations. A friendly graphical user interface （GUI） is designed for deployment of the software. Results： The preliminary study of the machine optimization shows that it is efficient to fred good pair of FODO lattice to get power high enough even if it is not globally optimized. Conclusion： This application of genetic algorithm for optimization design of free electron laser is efficient and stable based on many experiments of multi-variable optimization problems, and it is helpful to the future application to the genuine FEL machine optimization.

关 键 词：实数编码 遗传算法 自由电子激光 

分 类 号：TL506[核科学技术—核技术及应用]