go directly to the contents
EUChinaGRID

application repository

  • name: Cosmological Parameter Constraints with the Combination of Different Probes
  • domain: Astronomy
  • country: China
  • author:
  • institute: Department of Astronomy, Peking University
  • contacts:
  • description: To constraint the underlying cosmology, currently there are several promising approaches, including Cosmic Microwave Background, Supernova Ia, baryon acoustic oscillation and weak lensing. Although they are believed to be useful, each of them individually cannot impose tight enough constraints on cosmological parameters, since all of them will suffer from different parameter degeneracies. However, with the combination of some of them, the degeneracy can be break as different probes show different directions of degeneracy in the parameter space. We are planning to use grid to do the multi-probe fitting to some existing data, as well as some theoretical predictions for parameter constraining level of future surveys, which may work with one or several of the above probes.
    The application is a simple chi-square fitting process with Minuit (in CERNLIB), but the calculation of the theoretical model predictions to fit data is a challenging CPU intensive job, since we will combine several different cosmological probes, including supernova, cosmic microwave background radiation, weak lensing and galaxy power spectrum. Most of these probes should provide a big dataset with correlation inside to be fit. In addition, the datasets from different probes are usually correlated too, so we have to do the calculation of theoretical predictions for different probes all together. That is time consuming too besides the chi-square fitting for 8 or more cosmological parameters simultaneously.
    This application in astronomy field is connected with Zen Project [Zen], which link researchers from France and Beijing in solving dark matter enigma. Using EUChinaGRID infrastructure the researchers were constructing a Confidence Level contour (which takes from 2k to 38k CPU/hours depending on parameters set).
  • functionalities:
  • middleware requirements:
  • resources requirements:

Back to list of applications >>

powered by
Consortium GARR