Mugdim Bublin

A Cost-Function-Based Dynamic Channel Allocation and Its Limits

Design of efficient dynamic channel allocation (DCA) algorithms has an important impact on performance optimization of wireless mobile networks. The previously proposed DCA algorithms are shown to have good performance but only for certain services and under certain load conditions. Within this paper, we propose a new cost-function-based DCA algorithm, which is fully adaptive to service and load conditions and includes some well-known DCA algorithms such as random, minimum interference, and autonomous reuse partitioning DCA as special cases. By using dynamic system-level simulations, we show that the algorithm provides the same or substantially better performance than these DCA algorithms, with low signalization and computation overhead. We also show the general limits of DCA algorithms in the presence of other interference reduction techniques such as power control and smart antennas.

Inter-cell Interference Management by Dynamic Channel Allocation, Scheduling and Smart Antennas

An impact of dynamic channel allocation, opportunistic scheduling and smart antennas on inter-cell interference, and consequently on system performance is studied in this paper. It is shown that from the receiver point of view the most effective inter-cell interference reduction method is the opportunistic scheduling, followed by smart antennas, and dynamic channel allocation. Furthermore, interesting trade-offs that can be made by combining these techniques are observed as well.

A Self-Adaptive, Utility-based Scheduling for Wireless Cellular Networks

Design of efficient scheduling algorithms has an important impact on performance of wireless mobile networks. The previously proposed scheduling algorithms like Best CQI and Proportional Fair have good performance, but only for certain load and services types. Within this work, we propose a novel Utility based Scheduling algorithm, which is fully adaptive to service and load conditions and includes some well known scheduling algorithms as special cases. We show by using dynamic system level simulations that the algorithm provides same or substantially better performance than these scheduling algorithms, with low signalization and computation overhead.