论文标题
Millenniumtng项目：通过加速N体缩放和subhalo丰度匹配的星系聚类推断宇宙学
The MillenniumTNG Project: Inferring cosmology from galaxy clustering with accelerated N-body scaling and subhalo abundance matching
论文作者
论文摘要
我们介绍了一种新的技术，用于使用星系的非线性红移空间聚类来限制宇宙学参数和星系组装偏置。我们将宇宙学n体仿真和插入星系缩放，并以次荷兰的丰度匹配扩展（羞耻）经验模型生成超过175,000个跨越所有相关宇宙学和耻辱参数值的聚类测量值。然后，我们构建一个能够在单极，Quadrupole和Hexadecapole级别复制预计的星系相关功能的模拟器，用于分离$ 0.1 \，H^{ - 1} {\ rm Mpc} $和$ 25 \，$ 25 \，H^{ - 1} { - 1} {\ rm mpc} $。 We test this approach by using the emulator and Monte Carlo Markov Chain (MCMC) inference to jointly estimate cosmology and assembly bias parameters both for the MTNG740 hydrodynamic simulation and for a semi-analytical galaxy formation model (SAM) built on the MTNG740-DM dark matter-only simulation, obtaining unbiased results for all cosmological parameters.例如，对于MTNG740，星系号密度为$ n \ sim 0.01 H^{3} {\ rm mpc}^{ - 3} $，我们获得$σ_{8} = 0.799 = 0.799^{+0.039} {+0.039} 0.8159）和$ω_\ mathrm {m} h^2 = 0.138^{+ 0.025} _ { - 0.018} $（$ω_ {\ Mathrm {M}}对于固定的Hubble参数（$ H $），约束变为$ω_\ MATHRM {M} H^2 = 0.137^{+ 0.011} _ { - 0.012} $。对于SAM和其他测试的样品密度，我们的方法的性能类似。我们几乎总是恢复一个Sigma内的真实量的星系组件偏置。当包括小于$ 2 \，h^{ - 1} {\ rm mpc} $的尺度时，以及至少包含投影相关函数和单调时，可以获得最佳约束。这些方法提供了一种使用星系调查来限制宇宙参数的有力方法。
We introduce a novel technique for constraining cosmological parameters and galaxy assembly bias using non-linear redshift-space clustering of galaxies. We scale cosmological N-body simulations and insert galaxies with the SubHalo Abundance Matching extended (SHAMe) empirical model to generate over 175,000 clustering measurements spanning all relevant cosmological and SHAMe parameter values. We then build an emulator capable of reproducing the projected galaxy correlation function at the monopole, quadrupole and hexadecapole level for separations between $0.1\,h^{-1}{\rm Mpc}$ and $25\,h^{-1}{\rm Mpc}$. We test this approach by using the emulator and Monte Carlo Markov Chain (MCMC) inference to jointly estimate cosmology and assembly bias parameters both for the MTNG740 hydrodynamic simulation and for a semi-analytical galaxy formation model (SAM) built on the MTNG740-DM dark matter-only simulation, obtaining unbiased results for all cosmological parameters. For instance, for MTNG740 and a galaxy number density of $n\sim 0.01 h^{3}{\rm Mpc}^{-3}$, we obtain $σ_{8}=0.799^{+0.039}_{-0.044}$ ($σ_{8,{\rm MTNG}} =$ 0.8159), and $Ω_\mathrm{M}h^2= 0.138^{+ 0.025}_{- 0.018}$ ($Ω_{\mathrm{M}} h^2_{\rm MTNG} =$ 0.142). For fixed Hubble parameter ($h$), the constraint becomes $Ω_\mathrm{M}h^2= 0.137^{+ 0.011}_{- 0.012}$. Our method performs similarly well for the SAM and for other tested sample densities. We almost always recover the true amount of galaxy assembly bias within one sigma. The best constraints are obtained when scales smaller than $2\,h^{-1}{\rm Mpc}$ are included, as well as when at least the projected correlation function and the monopole are incorporated. These methods offer a powerful way to constrain cosmological parameters using galaxy surveys.
