We utilize a two-gas model to simulate collective oscillations of a Bose-Einstein condensate at finite temperatures. The condensate is described using a generalized GrossPitaevskii equation, which is coupled to a thermal cloud modelled by a Monte Carlo algorithm. This allows us to include the collective dynamics of both the condensed and non-condensed components self-consistently. We simulate quadrupolar excitations, and measure the damping rate and frequency as a function of temperature. We also observe revivals in condensate oscillations at high temperatures, and in the thermal cloud at low temperature. Extensions of the model to include non-equilibrium effects and describe more complex phenomena are discussed.PACS numbers: 03.75. Fi, 05.30.Jp, 67.40.Db The first experimental observation of Bose-Einstein condensation (BEC) in magnetically trapped alkali atoms in 1995 [1-3] was a precursor to an explosion of interest in the properties of weakly-interacting Bose gases. Much of the subsequent theory [4] has focused on the dynamics of the condensate, including phenomena such as collective excitations and vortex motion. In the limit of zero temperature, one can represent the condensate by a macroscopic wavefunction analogous to a classical field. In this case the behavior can be described in terms of the GrossPitaevskii (GP) equation, which has the form of a nonlinear Shrödinger equation. Extension of the description to finite temperatures, where one must include fluctuations upon the condensate wavefunction, is a considerable challenge. However, the motivation is clear, as such a description would allow direct comparison with experiments where a non-condensed thermal cloud is present, as well as revealing new phenomena such as damping of collective modes [5][6][7][8][9][10] and the decay of metastable vortices [11,12].Amongst the most compelling evidence for the validity of the GP equation at low temperatures is its quantitative agreement with experiment for low-energy collective modes. However, consistent theoretical descriptions at higher temperatures have proved far more elusive, where experiments have demonstrated marked frequency shifts and damping of the condensate modes in the presence of a significant non-condensed component [9,10]. Theoretical studies have tended to concentrate on one of two regimes, depending upon the density and temperature of the system. At high densities, where collisions are sufficiently rapid to force the system into local equilibrium, the dynamics of the condensate and thermal cloud can be described by a set of coupled hydrodynamical equations [13][14][15]. Damping mechanisms in this case are of a dissipative type (i.e. viscosity and thermal relaxation). For very dilute systems or at low temperatures the mean free path of the elementary excitations become comparable to the size of the system and collisions play only a minor role. Damping in this collisionless regime is not related to thermalization processes but to coupling between excitations, and can be described within the fr...