Emergence of region-based transmission when computation is unconstrained

Here, we study the emergence of a region-based approach when transmission costs, rather than computation constraints, limit the information upon which decisions are conditioned. We obtain that the average long-run loss of the transmission problem of size n is greater than that of the transmission problem of size Kn following a region-based approach, when computation is unconstrained. Hence a transmission problem of size Kn can achieve average costs lower than those of a transmission problem of size n by dividing the image into K quantizers of equal size n that imitate the prioritization protocol of an image of size n. In this case we have that additive and symmetric transmission costs, linearity and monotony of long-run loss, existence of cost-minimizing prioritization protocols, symmetric joint distribution of processes, not perfectly correlated processes, are among others some of the robust properties of constraints that drive the emergence of region-based transmission.