1: \begin{abstract}

2: The montecarlo method, which is quite commonly used

3: to solve maximum entropy problems in statistical physics, can actually

4: be used to solve inverse problems in a much wider context.

5: The probability distribution which maximizes entropy can be

6: calculated analytically by introducing Lagrange parameters.

7: The problem of fixing these lagrangean parameters is circumvented

8: by introduction of a microcanonical ensemble which describes

9: a system together with its heath bath. Some further simplifying

10: assumptions make it feasible to do montecarlo sampling

11: of the probability distribution.

12: The method is applied to the example of determining

13: the distribution of the density of the earth from three data.

14: Advantages of the method are guaranteed convergence

15: and a clear information-theoretic foundation.

16: 

17: \end{abstract}

18: