Because of their complexity models of supernova explosions have always been a challenge. In the case of core collapse models, which utilize gravitational binding for the explosion, the problem is not so much an energy but rather a momentum problem. There is plenty of energy available and only a small fraction of it has to be converted into outward momentum of the stellar envelope, but this is a non-trivial problem. Thermonuclear explosion models, on the other hand, suffer from the fact that once nuclear burning is ignited the progenitor star, presumably a white dwarf, it tends to expand and cool. It has been shown that an explosion results only if the burning front propagates with a velocity much larger than the laminar speed of nuclear flame in degenerate matter. Since for most supernova models hydrodynamic instabilities play a key role numerical experiments to prove or disprove certain ideas have to be 2- or 3-dimensional and, in fact, they have become feasible recently. Here we discuss the present status of 2- and 3-d numerical simulations of supernova explosions. In the case of explosions triggered by gravitational energy release, the crucial problem is the transport of neutrinos in convectively unstable matter, and for explosions powered by nuclear energy release, the problem of turbulent combustion has to be addressed.