Joint optimization of content selection and multicast grouping for heterogeneous requests

Develop methods to jointly optimize content selection and multicast grouping in wireless multicasting systems where users have heterogeneous (possibly overlapping) content requests, so that the resulting decisions effectively serve diverse user interests under practical wireless constraints.

Background

Most existing wireless multicast resource-allocation methods assume that all users request identical content, which does not hold in vehicle-to-infrastructure collaborative perception and similar applications where users have heterogeneous yet overlapping information needs. Broadcast and unicast approaches are inefficient in this setting because they either transmit irrelevant data to many users or fail to exploit shared interests.

Only a few prior studies consider heterogeneous interests, and those typically rely on heuristics without theoretical guarantees. Consequently, jointly deciding which content to transmit and how to form multicast groups when user requests differ has been identified as an open challenge motivating the development of Birdcast.