Is it acceptable to base a physical theory on a philosophical idea? Are theories that are based on the philosophy acceptable in Physics?
If the theory is based on the philosophy that obtained from laboratory experiments, can it be accepted?
 A: If a theory is not built on a solid foundation of (semi-rigorous) mathematics and a well-defined physical idea, the chances of it being accepted by the majority of physicists as a valid physical model are extremely small. If one wants to build up a theory of physics purely from philosphy, one will into some significant problems. After all is said and done, a theory must come up with some quantitative predictions for it to be any good.
However, this is not to say that philosophy has no role or place in physics. In fact, I think that the currently reasonably popular view that philosophy has nothing useful to contribute to physics is not only a philosophical statement in and of itself (and therefore immediately inconsistent) but also false. Whether we like it or not, physicists, too, are humans. When we think about theories and models of physical phenomena, it is really inevitable that our ideas, judgments and inspirations are partially due to things that are not strictly objective or scientific. This is not a problem and there is no need to hide it, but it is really unreasonable to deny it
One example that I think demonstrates my point quite well is the modern perspective of all physical theories as effective theories that should not expected to be absolutely true, but just capture the essential phenomena at some relevant (energy) scale of interest. Another example that is already being pointed out by Floris is the development of quantum mechanics (and relativity theory) in the early 20th century, which heavily relied on thought experiments that were often quite philosophical in nature.
A: If a theory never connects with experimental reality then no matter how neat it is it will eventually be dropped. This is the problem that much of theoretical physics based around the various incarnations of String Theory finds itself.
A theory is in one very real sense a data compression and predictive algorithm. It needs to join the experimental dots in the most efficient manner and also tell people where other dots might be found. Whether those "dots" are joined in a manner that creates a picture understandable by Humans seems (at present) to be irrelevant.
A: 
Are theories that are based on the philosophy acceptable in Physics?

Sure, but only if backed by maths and experiment. The prime example would be General Relativity, the development of which was guided by a whole bunch of principles (Mach's principle, equivalence principle, principle of covariance).
In contrast, Einstein failed to develop a unified classical field theory, arguably because there was no new guiding principle and he ended up just toying with the mathematics. (Note that Einstein actually did end up with a candidate theory that he hadn't discarded as of his death).
String theory as of today is probably in better shape than classical unification was back then, but it hasn't been universally accepted precisely because the most important ingredient besides philosophy and mathematics is still missing: experimental verification.
