The motion of artificial microswimmers can strikingly resemble collective motion in biological systems even though it only involves physical and chemical processes. A detailed understanding of their emergent swarming properties may therefore help to distinguish merely physics-related from biology-related aspects of motion in biological systems. This will improve our understanding of biological microswimmers, such as bacteria and spermatozoa, and it will provide the background to target the design of swimmers for technical and medical applications.Here, we will focus on modelling a new type of artificial microswimmers where propulsion is achieved by Marangoni flow. To fully characterize the parameter dependence of their motion, we will develop a multiscale description addressing (i) the propulsion mechanism of single swimmers, (ii) interactions between small numbers of swimmers, and (iii) the collective behaviour of large assemblies. We will derive the flow field inside and outside individual droplets, and take into account the two-way coupling of the swimmers motion and the external flow on all the modelling levels. For small numbers of droplets we will develop a flexible CFD model based on a level set method, easily adaptable to different kinds of swimmers. This will allow us to explore the dependence of propulsion on experimentally tunable parameters, like the droplet size and surfactant concentration, and to fully characterize the interactions between swimmers. To simulate the collective behaviour L. Stricker will develop a point-particle combined Lagrangian-Eulerian model that solves the flow and models each swimmer as a rigid particle, undergoing the forces established by the level-set simulations. For such code she will import state-of-the-art techniques from simulations of inertial particles and bubbly flows. All her results will be compared to predictions of theoretical models and to experimental data collected at the MPI-DS.