The formation of massive stars yields strong feedback effects onto its host core and stellar cluster environment via protostellar outflows, radiation heating, radiation pressure, ionization, stellar winds and mass loss, and supernova (in chronological order). From a theoretical point of view, the question arises, how a massive protostar is able to accrete its mass up to the observed upper mass limit despite its strong radiation pressure feedback. In this talk, I present results of several series of self-gravitating radiation-hydrodynamics simulations of core collapse towards high-mass star formation, including the effects of radiation heating, radiation pressure, protostellar outflows, and ionization. We propose a solution to the radiation pressure problem in the formation of massive stars via the so-called flashlight effect: due to the formation of an optically thick accretion disk, the thermal radiative flux becomes strongly anisotropic and preferentially escapes through the disk's atmosphere, i.e. perpendicular to the sustained accretion through the disk's mid-plane. Furthermore, including feedback of early protostellar outflows yields a large scale anisotropy, which extends the disk's flashlight effect from the few hundred AU scale of the circumstellar disk to a core's flashlight effect up to a 0.1-parsec scale. This core’s flashlight effect allows core gas to accrete on the disk for longer, in the same way that the disk’s flashlight effect allows disk gas to accrete on the star for longer. In summary, I will demonstrate straight-forward mechanisms, which allows the formation of the most massive stars known in the present-day universe despite of their strong feedback via outflows and radiation. The basic theory of star formation herein is just a scaled-up version of low-mass star formation, i.e. based on accretion from core to disk to protostellar scales.
