This talk is devoted to a mathematical analysis of the time reversal
method which was promoted by Mathias Fink, his group and others.

It involves source and transductors. The challenge is to
understand how to use as few transductors as possible.

Emphasis is put on examples of problems in a closed bounded cavity In
this situation I will describe the effect of ergodicity both when the
transductors are in the media or when they are at the boundary.

Results and methods are compared with what has already been done for
random media. Some of the proof are very similar sharper results are
obtained but only in domains with no boundary.