Interacting Urn Models

The aim of this paper is to study the asymptotic behavior of strongly reinforced interacting urns with partial memory sharing. The reinforcement mechanism considered is as follows: draw at each step and for each urn a white or black ball from either all the urns combined (with probability $p$) or the urn alone (with probability $1-p$) and add a new ball of the same color to this urn. The probability of drawing a ball of a certain color is proportional to $w_k$ where $k$ is the number of balls of this color. The higher the $p$, the more memory is shared between the urns. The main results can be informally stated as follows: in the exponential case $w_k=\rho^k$, if $p\geq 1/2$ then all the urns draw the same color after a finite time, and if $p<1/2$ then some urns fixate on a unique color and others keep drawing both black and white balls.