Collaborative Research: CDS&E: Mining Physically Predictive Cosmological Simulations

通过运行详细的计算机模拟，研究人员正在努力了解星系是如何形成的。由于这个问题的复杂性，需要大型计算机模拟来跟踪所有关键的物理过程。这些研究人员正在开发一种高效建模的新方法。事实证明，新方法在再现许多观测到的星系特性方面非常成功。研究人员将进一步改进他们的模拟，以研究更广泛的星系类型。他们将通过包括更多的物理过程和提高数值精度，在多个方面推动星系形成模型的前沿。新的模拟将用于研究不同大小的星系，从矮星系到比我们银河系大10倍的星系晕。研究人员将训练学生掌握具有广泛科学和工业应用的计算机建模技术。他们还将根据新的模拟制作电影质量的动画，这将成为教育和推广活动的基础，包括几个天文馆。模拟代码的新模块和优化将公开发布，这不仅有利于天体物理学，也有利于其他学科，如工程。研究人员领导了“现实环境中的反馈”（FIRE）项目，并创建了一系列星系形成的宇宙学模拟，明确地在单个恒星形成区域的尺度上模拟恒星反馈。在这些模拟中，反馈将直接锚定在恒星演化模型上，通过减少对可调参数的依赖，显著提高了它们的预测能力。FIRE模拟在匹配恒星质量、星系风和矮星系暗物质分布方面已经被证明是成功的。为了吸引更多的观众，研究人员计划为FIRE模拟创造新的音乐“声音”。这些声音将在一年一度的音乐会上表演，并强调与时间有关的现象。

By running detailed computer simulations, the investigators are working to understand how galaxies form. Because of the complexity of this problem, large computer simulations are needed to follow all the key physical processes. These investigators are developing a new approach for efficient modeling. The new methods have proved extremely successful in reproducing many of the observed properties of galaxies. The investigators will further improve their simulations to study a much wider range of galaxy types. They will push the frontiers of galaxy formation modeling on multiple fronts by including more physical processes and improving numerical accuracy. The new simulations will be used to study different sized galaxies from dwarf galaxies to galaxy halos ten times more massive than our Milky Way galaxy. The investigators will train students in computer modeling techniques that have a wide range of scientific and industrial applications. They will also produce film-quality animations produced from the new simulations, which will be the basis for education and outreach initiatives, including several planetariums. The new modules and optimizations for the simulation code will be publicly released, which will benefit not only astrophysics but also other disciplines, such as engineering.The investigators lead the "Feedback In Realistic Environments" (FIRE) project and create sets of cosmological simulations of galaxy formation that explicitly model stellar feedback on the scale of individual star-forming regions. In these simulations feedback will be directly anchored to stellar evolution models, significantly improving their predictive power by reducing the reliance on adjustable parameters. The FIRE simulations have proven successful in matching stellar masses, galactic winds, and the dark matter distributions of dwarf galaxies. To reach larger audiences, the investigators plan to create new musical "sonifications" of the FIRE simulations. The sonifications will be performed at an annual concert and highlight time-dependent phenomena.