论文部分内容阅读
捅要:为了最大化动态微蜂窝网络(SCNs)的全网吞吐量,并同时解决微蜂窝网络中微蜂窝基站状态动态性的难点,文章基于博弈论和传统随机学习算法,提出一种动态非对称图博弈模型,以及一种动态联合信道功率选择随机学习算法DJCPS-SLA来研究并解决动态活跃基站下的联合信道功率分配问题。前人提出的优化算法不能同时优化信道和功率两个决策,也没有考虑微蜂窝网络的动态性。因此,解决这类动态联合优化问题是有挑战性的,特别是活跃基站集合在时隙间会动态变化的这种情况。DJCPS-SLA的仿真结果佐证了算法的收敛性,结果表明该算法具有较快的收敛速度,并且其吞吐量性能远远超过随机选择算法。
In order to maximize the throughput of dynamic micro-cellular networks (SCNs) and solve the difficulties of the dynamic state of micro-cellular base stations in micro-cellular networks at the same time, this paper proposes a dynamic non-dynamic learning algorithm based on game theory and traditional random learning algorithm Symmetric graph game model and a dynamic joint channel power selection random learning algorithm DJCPS-SLA to study and solve the joint channel power allocation problem under dynamic active base station. The optimization algorithm proposed by our predecessors can not optimize both the channel and the power simultaneously, and does not consider the dynamics of the micro-cellular network. Therefore, it is challenging to solve this kind of dynamic joint optimization problem, especially when the active base station aggregation changes dynamically between timeslots. The simulation results of DJCPS-SLA demonstrate the convergence of the algorithm. The results show that the proposed algorithm has a faster convergence rate and its throughput performance far exceeds the random selection algorithm.