Three novel opportunistic scheduling algorithms in CoMP-CSB scenario

Abstract

Coordinated scheduling/beamforming (CSB), which belongs to the coordinated multi-point (CoMP) transmission, has received lots of attention recently due to its great potential to mitigate inter-cell interference (ICI) and to increase the cell-edge throughput, and meanwhile it only requires limited base station cooperation and is easy to implement. However, to the best of our knowledge, there are no effective scheduling algorithms with low complexity and overhead in CoMP-CSB scenario as yet. Thus, in this paper, we propose three novel opportunistic scheduling algorithms in CoMP-CSB scenario. All of them jointly consider the intended channel condition of the scheduled user from its serving cell and the orthogonality between the intended channel and the corresponding interference channels to concurrently scheduled users in nearby cells, thus exploiting multi-user diversity (MUD) and mitigating ICI at the same time. Algorithm 1 cooperatively chooses the most orthogonal user pair within a candidate user set in which all users have a large local channel feedback, while Algorithm 2 concurrently schedules the user pair with the largest ratio between the local channel feedbacks and the aforementioned orthogonality within the same candidate user set. Algorithm 3 performs in the way similar to the proportional fairness scheduling, while making a proper modification for its usage in CoMP-CSB scenario. The performance of the proposed scheduling algorithms are evaluated through simulation. Results show that, they all can significantly enhance the received signal to interference plus noise ratio (SINR) with relatively good fairness guarantee, thus achieving larger throughputs and utilities than several well-known scheduling algorithms. Algorithm 2 even outperforms Algorithm 1 when the aforementioned candidate user set is big enough in size and has a bit more overhead/complexity. Furthermore, Algorithms 3 is the best one among all the three proposed algorithms, but it requires more overhead/complexity than Algorithm 1 and 2. Finally, we give the optimal parameter for all of the three proposed algorithms, which can make a good tradeoff between performance and overhead/complexity.