中文摘要：

频谱资源异构性是5G通信系统的重要特性之一。为实现频谱资源分配与需求的最优匹配，关注5G微蜂窝的异构信道选择问题。针对该问题，传统集中式优化机制系统效率较优但优化开销大，而传统分布式优化机制优化开销较少但系统效率受限。为实现系统效率与优化开销的有效折中，将优化问题建模为局部合作博弈，提出基于局部信息交互的博弈学习算法，实现了系统在分布式优化机制下的最优频谱资源分配。仿真结果验证了算法的最优性、收敛性和稳健性。

英文摘要：

Spectral heterogeneous is one of the most important characteristics in 5G communication system. To achieve the optimal match between resource allocation and resource demand, this paper focused on the heterogeneous channel selection problem for 5G small cells. The traditional centralized optimization scheme deserved high system efficiency but also high overhead, while the traditional distributed optimization scheme had low overhead but also low system efficiency. To achieve the optimal tradeoff between system efficiency and optimization overhead, this paper formulated the problem as a local cooperation game, proposed the game learning algorithm based on the local information exchange, and realized the optimal resource allocation strategy in a distributed manner. Simulation results validate the optimality, convergence and stability of the proposed algorithm.