A Monte Carlo simulation-based optimization of a multilayer 10B-RPC thermal neutron detector is performed targeting an increase in the counting rate capability while maintaining high (>50%) detection efficiency for thermal neutrons. The converter layer thicknesses of individual RPCs are optimized for several configurations of a detector containing a stack of 10 double gap RPCs. The results suggest that it is possible to reach a counting rate which is by a factor of eight higher in comparison to the rate of a detector with only one double-gap RPC. The effect of neutron scattering inside the detector contributing to the background is analyzed and design modifications of the first detector prototype, tested at neutron beam, are suggested.