This website requires JavaScript.

Bayesian optimization of laser-plasma accelerators assisted by reduced physical models

A. Ferran PousaS. JalasM. Kirchen ...+6 R. Lehe
Dec 2022
摘要
Particle-in-cell simulations are among the most essential tools for themodeling and optimization of laser-plasma accelerators, since they reproducethe physics from first principles. However, the high computational costassociated with them can severely limit the scope of parameter and designoptimization studies. Here, we show that a multitask Bayesian optimizationalgorithm can be used to mitigate the need for such high-fidelity simulationsby incorporating information from inexpensive evaluations of reduced physicalmodels. In a proof-of-principle study, where a high-fidelity optimization withFBPIC is assisted by reduced-model simulations with Wake-T, the algorithmdemonstrates an order-of-magnitude speedup. This opens a path for thecost-effective optimization of laser-plasma accelerators in large parameterspaces, an important step towards fulfilling the high beam quality requirementsof future applications.
展开全部
图表提取

暂无人提供速读十问回答

论文十问由沈向洋博士提出,鼓励大家带着这十个问题去阅读论文,用有用的信息构建认知模型。写出自己的十问回答,还有机会在当前页面展示哦。

Q1论文试图解决什么问题?
Q2这是否是一个新的问题?
Q3这篇文章要验证一个什么科学假设?
0
被引用
笔记
问答