Bachelor Thesis With Prof. Lesgourgues in 2023


1. The early universe explored with the C++ simulation code CosmoLattice

Context: CosmoLattice is a recent public code for lattice simulations of the evolution of fields in an expanding universe. It can be used to simulate interesting phenomena that might have occured in the early universe, like a stage of cosmological inflation, some phase transitions, or the production of gravitational waves. The RWTH cosmology group has expertise on the underlying theory and on several types of cosmology simulation codes, but not on this particular type of code. However, we would like to develop new research project based on the use of CosmoLattice.

Goals: The student will pave the way and learn how to intsall the code, how to run it (either on their laptop or on the RWTH cluster), how to study simple models, and how to visualize the results in a nice way (with colourful images or videos). We have close contacts with the authors of the code, who will be ready to help on request.

Skills required: numerical skills (mainly in C++ and python), interest for numerical simulations and for theoretical physics (field theory).

Skills learned: practise with advanced scientific software, some knowledge of the physics of the early unvierse

Main supervisor: Prof. Julien Lesgourgues

2. The nonlinear Universe explored with the python N-body code CONCEPT

Context: Our Universe became very clumpy due to the gravitational interactions which cause, among others, the appearance of the cosmic web structure at large scales. To study this structure, cosmologists have developed N-body simulation codes, that track the evolution of billions of particles and their gravitational interactions. So far, these simulations have been coded in programming languages that are difficult to read and maintain. Recently, an N-body code called CONCEPT has been released, which is computationally efficient and at the same time, easy to understand.

Goals: In this project, the student will learn how to run and use the code in the RWTH cluster. They will produce data that can be then emulated and tested against cosmological high-precision data coming from future missions, such as Euclid. Optionally, the student would also learn how to visualize the output of N-body simulations, in order to produce colorful images and videos.

Skills required: programming skills, numerical methods, and to a lesser degree statistical methods for data analysis.

Skills learned: advanced python coding, working and understanding N-body simulations, analyzing numerical data, cosmology of the large scale structure and the cosmic web

Main supervisor: Dr. Santiago Casas

3. Data-driven solutions for varying electron masses in the early universe with genetic algorithms

Context: Despite the overall success of the cosmological standard model, a significant tension has emerged between different measurements of the expansion rate of the universe. This might be a hint for new physics. It has been shown that time-varying electron mass in the early universe can resolve this tension - while being compatible with some fundamental physics models. Yet, we need to infer from cosmological data the phenomenological behavior that this variation might follow.

Goals: The student is expected to learn about this tension and the possibility to resolve it with a time-varying electron mass. Then, they will work with a modified version of a cosmology simulation code. They will use a genetic algorithm to test and find different analytical forms of this variation. The quality of a function will be tested upon real observations. This may help future cosmologists to find proper models which intrinsically explain the mass variation.

Skills required: programming skills (mainly in python), interest in machine learning and in cosmology.

Skills learned: advanced python coding, experience in machine learning techniques, analyzing numerical data, understanding of cosmological observations

Main supervisor: MSc. Sven Guenther