Exponential stochastic compression of one-dimensional space and 146 percent

Exponential stochastic compression is the process when every second cell of an infinite chain may increase its weight merging randomly with left, right, or both neighboring cells. The total mass conservation is assumed. After that, merged cells fill the empty space, compressing the chain twice. They may fill empty spaces in two different ways: (I) using shifts only, i.e. preserving the order; (II) using shifts and random permutations. Compressing the initial homogeneous chain with cell weights $1$ many times, we compute final densities ${\rho }_i$ of cells with weight $i=1,2,3,\mathrm{.}\mathrm{.}\mathrm{.}$. The main result is that ${\rho }_i\mathrm{/}{\rho }_1=i$ in the ordered case (I), and ${\rho }_i\mathrm{/}{\rho }_1\approx \mathrm{1.464910...}(i-1\mathrm{/}4)$ in the disordered case (II). The multiplier in the disordered case has a fractal nature. The compression of initially inhomogeneous chains and rescaled continuous densities are also discussed.