Self-adaptive Genetic algorithm based MU-MIMO scheduling scheme

Multi-user Multiple Input Multiple Output (MU-MIMO) system is known to enhance the system capacity with low network delay. One of the biggest challenges with MU-MIMO is on the scheduling scheme which simultaneously selects multiple users to maximize the sum rate. Genetic Algorithm (GA) works perfectly as an optimal or suboptimal solution with quite low complexity to handle such problems. In this paper, we promote a MU-MIMO scheduling scheme which is based on a modified GA, namely Self-adaptive GA (SaGA). SaGA can adjust the performance of population and generation dynamically during run. Furthermore, a self-adaptive elitism strategy also makes contributions to a better behave and faster convergence beyond the traditional GA strategy. Simulation results demonstrate that the performance of SaGA based scheduling is quite approximate to the exhaustive searching while with a much lower complexity.