A mean-field approximation for the diffusion coefficient in lattice Lorentz
gases with an arbitrary mixture of pointlike stochastic scatterers in the
low-density limit is proposed. In this approximation, the diffusion coeffic
ient is directly related to the first return probability of the moving part
icle in the corresponding Cayley tree through an effective ring operator. A
renormalization scheme for the approximate determination of the first retu
rn probability is constructed. The predictions of this mean-field theory an
d those of the repeated ring approximation (RRA) are compared with computer
simulation results for models in which a fraction x(B) of the scatterers a
re deterministic backscatterers, so that the diffusion coefficient vanishes
beyond a certain percolation threshold x(B)(c),. The approximation propose
d in this paper is seen to be in good agreement with the simulation results
, in contrast to the RRA, which already fails to give the correct percolati
on threshold. [S1063-651X(99)10807-9].