基于遗传算法求解三维匹配的资源分配问题

Improved genetic algorithm for solving resource allocationproblem of 3D matching

针对非正交多址接入(NOMA)技术的两层异构网络(HetNets)的资源配置,因用户、基站和子信道三维匹配属于NP难题,多分解为二维匹配求解,为此提出一种改进的遗传算法(GA)求解用户的多维匹配。为满足系统总容量最大并降低时间复杂度,将遗传算法的编码方式设计为一种多维映射过程;为防止陷入局部最优并提高全局搜索能力,对选择算子进行确定性和随机性的结合。实验结果表明,该算法相对于贪婪算法和双边匹配算法,具有收敛速度快和全局性更好等优点。

Aiming at the resource allocation problem of the two-layer heterogeneous network(HetNets)of non-orthogonal multiple access(NOMA)technology,because the three-dimensional matching of users,base stations and sub-channels is an NP problem,the three-dimensional matching is decomposed into two-dimensional matching for solving in general.To solve this problem,an improved genetic algorithm(GA)was proposed.To maximize the total system capacity and reduce the time complexity,the coding method of the genetic algorithm was designed as a multi-dimensional mapping process.To prevent the premature convergence problem and improve the global-optimization capability,the selection operator was combined with deterministic and random features.Results of the simulation indicate that the proposed GA has higher convergence speed and better optimization perfor-mance than greedy algorithm and two-sided matching algorithm.