Distributed Constrained Steiner Tree Algorithm for Multicast in Ad hoc Networks

The paper formally defines the problem of constrained Steiner tree (CST) in the context of wireless ad hoc networks. In contrast to wired networks, due to the broadcast nature of wireless communications, the notion of cost and delay is defined based on node, rather than link. By adapting the metaphor of Swarm Intelligence, the paper describes a distributed and asynchronous algorithm, termed CSTMAN, that constructs CSTs to facilitate multicast operations in ad hoc networks. Small control packets acting like ants are periodically generated by the group members to opportunistically discover better paths with lower total forwarding cost, while ensuring that delay (or hop count) constraints are not being violated. As a result, a CST keeps evolving into states that yield lower total forwarding cost in terms of transmission power, while satisfying the imposed hop count constraints. Simulation results, depicted in the forms of topology snapshots and forwarding cost variations over time, validate the effectiveness and reveal the characteristics of CSTMAN.