We confront the recent proposal of Emerging Brane Inflation with WMAP3+SDSS, finding a scalar spectral index of $n_s=0.9659^{+0.0049}_{-0.0052}$ in excellent agreement with observations. The proposal incorporates a preceding phase of isotropic, non accelerated expansion in all dimensions, providing suitable initial conditions for inflation. Additional observational constraints on the parameters of the model provide an estimate of the string scale. A graceful exit to inflation and stabilization of extra dimensions is achieved via a string gas. The resulting pre-heating phase shows some novel features due to a redshifting potential, comparable to effects due to the expansion of the universe itself. However, the model at hand suffers from either a potential over-production of relics after inflation or insufficient stabilization at late times.

We propose a new way of obtaining slow-roll inflation in the context of higher dimensional models motivated by string and M theory. In our model, all extra spatial dimensions are orbifolded. The initial conditions are taken to be a hot dense bulk brane gas which drives an initial phase of isotropic bulk expansion. This phase ends when a weak potential between the orbifold fixed planes begins to dominate. For a wide class of potentials, a period during which the bulk dimensions decrease sufficiently slowly to lead to slow-roll inflation of the three dimensions parallel to the orbifold fixed planes will result. Once the separation between the orbifold fixed planes becomes of the string scale, a repulsive potential due to string effects takes over and leads to a stabilization of the radion modes. The conversion of bulk branes into radiation during the phase of bulk contraction leads to reheating.

We propose a mechanism for solving the horizon and entropy problems of standard cosmology which does not make use of cosmological inflation. Crucial ingredients of our scenario are brane gases, extra dimensions, and a confining potential due to string gas effects which becomes dominant at string-scale brane separations. The initial conditions are taken to be a statistically homogeneous and isotropic hot brane gas in a space in which all spatial dimensions are of string scale. The extra dimensions which end up as the internal ones are orbifolded. The hot brane gas leads to an initial phase (Phase 1) of isotropic expansion. Once the bulk energy density has decreased sufficiently, a weak confining potential between the two orbifold fixed planes begins to dominate, leading to a contraction of the extra spatial dimensions (Phase 2). String modes which contain momentum about the dimensions perpendicular to the orbifold fixed planes provide a repulsive potential which prevents the two orbifold fixed planes from colliding. The radii of the extra dimensions stabilize, and thereafter our three spatial dimensions expand as in standard cosmology. The energy density after the stabilization of the extra dimensions is of string scale, whereas the spatial volume has greatly increased during Phases 1 and 2, thus leading to a non-inflationary solution of the horizon and entropy problems.

It is widely accepted that moduli in the mass range 10eV - $10^4$GeV which start to oscillate with an amplitude of the order of the Planck scale either jeopardize successful predictions of nucleosynthesis or overclose the Universe. It is shown that the moduli problem can be relaxed by making use of parametric resonance. A new non-perturbative decay channel for moduli oscillations is discussed. This channel becomes effective when the oscillating field results in a net negative mass term for the decay products. This scenario allows for the decay of the moduli much before nucleosynthesis and, therefore, leads to a complete solution of the cosmological moduli problem.