Well GR is not for sure the only gravitation theory which is feasible... there are a lot of modified theories which extend GR and which may be suitable as well, sice they reduce to GR at Solar System level while trying to explain more in other regimes. GR though is the oldest theory of this kind (all the modified gravity theories are much more recent) and helds anyway a lot of good predictions, so unless some of the extended model prove to be undoubtedly better than GR itself by explaining data GR couldn't, I guess everyone will go on working with GR.

Anyway the research on these alternative theories is quite active, even without thinking about quantum aspects.

The following is not endorsement of an alternative, but if someone wants an alternative model to compare with modern data there is one off-hand that I can name: Once upon a time, there was an alternative model -- this goes back about 40 years -- based on the conformal group. The theory and book were due to Segal. The assertion was that the model fitted cosmological data about as well as the standard model did. There was some argument, about the time I lost track, as to whether there was a significant disagreement with a particular set of measurements, or whether the measurements were actually on two somewhat different objects at different distances, leading to a set of one or two preferred curves resembling a not-equals sign. The conformal group universe was vastly older than the standard universe, avoiding the inflation issue that had been discovered at about the same time. This is not my current research interest, but the question was asked.

The reason is that the Einstein-Hilbert action, that contains the Ricci scalar and the cosmological constant term, is the most general action, that contains terms, with at most two derivatives and is invariant under general coordinate transformations (in four spacetime dimensions, of course)-and involves only the metric tensor.

If one is interested in the constraint that the *equations of motion* don't contain terms with more than two derivatives, then one *can* write down actions with more terms (e.g. the so-called ``Galileon'' terms). Such terms face challenges in order to satisfy tests at the Solar System level *and* to provide an alternative description of effects at cosmological scales-which is the reason they were introduced.

On my part I think that all our theories have a limited scope and are simplified representations of reality valid under certain conditions. This is what happened to Newton's theory of space and time and gravity.

Einstein's general relativity, or any other scientific theory is not exempt from this limitation, but broaden the horizon of our knowledge. For this reason it seems a mistake to attribute to this theory (or any other) the ability to provide an overview of all the infinite motion of matter.

Bearing this in mind, I propose to investigate the Relativistic Theory of Gravitation (RTG) of Anatoly Logunov. This theory extends Einstein's special relativity to accelerated systems and gravitational fields, understands gravity as a physical field, equal to the electric field, does not resort to the curvature of space to explain the distortions that causes gravity on the movement and attributed a different meaning to the mathematical building Minkoskwi, Ricci, etc, etc.

In my view, although it has some bugs, corrects essential difficulties of general relativity

The real problem with using GR to predict the initial singularity, or the initial origin if you will, is that this would be quantum mechanical as it is so small. GR cannot handle it. Indeed, how would you ever get an expansion it the mass starts out sitting inside its own event horizon, it would be a black hole.

Of course we can get around this by stating that there is no spacetime to distort. But then GR cannot handle this either. No spacetime? No Einstein-Hilbert Action.

The terms ``origin'' and ``fate'' of the Universe are sufficiently vague that it's not obvious how ``close'' to singularities the description applies. It's useful, in this regard, to read the paper of Hawking and Penrose, ``The Singularities of Gravitational Collapse and Cosmology'', http://rspa.royalsocietypublishing.org/content/314/1519/529

-it's extremely well-written.

As mentioned by Mattia Colombo, there are many theories that deal with the origin of the universe and its fate. But I like to point out that there is one theory that also deals with the presence of consciousness in the universe. It is "Periodic relativity" and "Periodic quantum gravity and cosmology".

Carlos, what do you mean with RTG "extends Einstein's special relativity to accelerated systems"? There has never been a problem with accelerated or other coordinate systems in SR, they are simply not preferred as inertial systems.

What do you think are the problems of RTG? IMHO, the main problem is the causality condition, which is not compatible with the equations: Even if initial values fulfill the causality condition, it may later be violated. What happens then? (The other one is a history of incorrect criticism of GR, but this is not a problem of the theory itself.)

This is an advantage of my ether variant of RTG http://arxiv.org/abs/gr-qc/0205035 - it does not need the RTG causality condition, replaces it by another one, which has a clear physical meaning: The ether density should be greater zero.