We study a data dissemination scenario in which data items are to be transmitted to mobile clients via one of the stationary data access points (APs) that the clients pass by en route to their destinations. The scheduler dedicates sequences of consecutive timeslots of an AP to downloading a data item to a client during the time window in which it is in range, which corresponds to assigning a job (the client's download) to a machine (the AP) among many. The transmission rate chosen for each assignment partly corresponds to setting a machine's speed, but it also has subtler effects. The APs may control transmission power to tune its transmission range making sure that no interference occurs with neighboring APs' transmissions. The problem is a generalization of an already NP-hard parallel-machine scheduling problem in which jobs' release times and deadlines depend on the machine to which they are assigned. We define this joint timeslot, power control, and rate assignment problem formally and apply both new algorithms and adaptations of existing algorithms to it. We evaluate these algorithms through simulations which show that our proposed algorithms achieve near-optimal throughput.
All Science Journal Classification (ASJC) codes
- Computer Science Applications
- Electrical and Electronic Engineering
- Applied Mathematics