基于遗传算法的最大化公平性基站主动存储方案

Maximum Fairness Basestation Active Storage Scheme based on Genetic Algorithm

提出了一个最大公平性存储方案(Maximum Fairness Storage Allocation Scheme,MFSA),以解决基站主动存储中资源分配不公平的问题。首先,对基站主动存储的公平性问题进行建模。具体来说,采用随机线性网络编码的方式将内容存储在基站,在总的存储量和传输延迟限制下,将存储分配方案建模为最大化小基站访问资源的Jain公平性。其次,设计可以求解问题的遗传算法。该遗传算法采用矩阵编码,利用罚函数法将有约束优化转换为无约束优化,同时利用模拟退火方法加速遗传算法的收敛速度。最后,利用设计的遗传算法对该问题进行求解。实验数据表明,与对比算法的存储方案相比,提出的存储方案在公平性上有较大提升,公平性指数在总的存储量限制为3 000、4 000和5 000时分别提高了17.01%、19.10%和18.20%,在传输延迟限制为50、60、70和80时分别提高了12.51%、15.10%、20.54%和21.20%。

MFSA （Maximum Fairness Storage Allocation） scheme is proposed to solve the problem of unfair resource allocation in basestation active storage. Firstly, the fairness problem of basestation active storage is modeled. Specifically, the content is stored in the basestation by means of random linear network coding. The storage allocation scheme is modeled to maximize Jain fairness of small basestation access resources under total storage and transmission delay constraints. Secondly, a genetic algorithm that can solve the problem is designed. And this genetic algorithm, by adopting matrix coding and using the penalty function method, transforms the constrained optimization into unconstrained optimization, and then by using the simulated annealing method, accelerates the convergence speed of the genetic algorithm. Finally, the problem is solved by using the designed genetic algorithm. The experiment indicates that compared with the storage scheme of the comparison algorithm, the storage scheme has a greater improvement in fairness. The fairness index is increased by 17.01%, 19.10%, and 18.20% when the total storage capacity being limited to 3000, 4000, and 5000, respectively; and increased by 12.51%, 15.10%, 20.54% and 21.20% when the transmission delay being limited to 50, 60, 70, and 80, respectively.